Focus measurement (more general image analysis question)
Locked
12
12
 
|
Hello,
sorry, the question is not that strongly related with ImageJ but maybe there is still someone who can help? I have a huge set of microscopic images from a confocal microscope (cells/nuclei + marker) and I'd like to automatically filter images which have a bad focus. First, I was thinking about some gradient values to judge but they are so dependent of the content (number of cells... , presence of the marker). Is there any reliable method known to measure and compare the focus of images??? Or does anybody know papers dealing with this issue??? (If there is something available in ImageJ it would be fine but I could not find anything...) Antje ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
You are tackling an interesting question. To be in focus is always
meant for an object, adjust focus is meant for a whole image. Hence the problem with focus is the observer, a cell of certain thickness might be in focus at the cell border or at the cell center. However, still there is a more important question. Images from confocal microscopes are by definition in focus of course related to the 'pin-hole' size. If you call such images in 'bad focus' maybe there is some help in microscope adjustment necessary? If you have stacks of images, images taken at different focal adjustments there is a program to merge such stacks into one image (e.g. plugin Exteded depth of field). However, my first remark is still valid, meaning the outcome of such a method is possibly far away from expectation. Regards Karsten Am 06.03.2007 um 09:17 schrieb Antje: > Hello, > > sorry, the question is not that strongly related with ImageJ but > maybe there is still someone who can help? > > I have a huge set of microscopic images from a confocal microscope > (cells/nuclei + marker) and I'd like to automatically filter images > which have a bad focus. First, I was thinking about some gradient > values to judge but they are so dependent of the content (number of > cells... , presence of the marker). Is there any reliable method > known to measure and compare the focus of images??? Or does anybody > know papers dealing with this issue??? > (If there is something available in ImageJ it would be fine but I > could not find anything...) > > Antje > > > ___________________________________________________________ > Telefonate ohne weitere Kosten vom PC zum PC: http:// > messenger.yahoo.de |
|
|
Hi Karsten,
you're right confocal microscopic images should be in focus by definition. It's for sure that our microscope needs to undergo improvements... Anyhow, due to a large screening mode of this microscope, I have to cope with out of focus images (unfortunately, I'm not that deep into physics to be able to explain the reason for this focus problems). I don't have stacks. Ciao, Antje Karsten Rodenacker schrieb: > You are tackling an interesting question. To be in focus is always meant > for an object, adjust focus is meant for a whole image. Hence the > problem with focus is the observer, a cell of certain thickness might be > in focus at the cell border or at the cell center. > > However, still there is a more important question. Images from confocal > microscopes are by definition in focus of course related to the > 'pin-hole' size. If you call such images in 'bad focus' maybe there is > some help in microscope adjustment necessary? > If you have stacks of images, images taken at different focal > adjustments there is a program to merge such stacks into one image (e.g. > plugin Exteded depth of field). However, my first remark is still valid, > meaning the outcome of such a method is possibly far away from expectation. > Regards > Karsten > > Am 06.03.2007 um 09:17 schrieb Antje: > >> Hello, >> >> sorry, the question is not that strongly related with ImageJ but maybe >> there is still someone who can help? >> >> I have a huge set of microscopic images from a confocal microscope >> (cells/nuclei + marker) and I'd like to automatically filter images >> which have a bad focus. First, I was thinking about some gradient >> values to judge but they are so dependent of the content (number of >> cells... , presence of the marker). Is there any reliable method known >> to measure and compare the focus of images??? Or does anybody know >> papers dealing with this issue??? >> (If there is something available in ImageJ it would be fine but I >> could not find anything...) >> >> Antje >> >> >> ___________________________________________________________ Telefonate >> ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
Re: Focus measurement (more general image analysis question)
|
|
Dear Antje,
I'd suggest to use an edge detection filter (e.g. Sobel filter "Find Edges") and to take the brightness of the result as a score. Images with sharp edges should produce a high score that permits you to sort your images. I don't know if this will work with your set of images, but it should be a quick way to get some idea of the "focus" quality (provided your objects are mainly in a single plane of focus). Best, Jan -------- Original Message -------- From: Antje <[hidden email]> To: [hidden email] Subject: Re:Focus measurement (more general image analysis question) Date: 06.03.2007 10:04 > Hi Karsten, > > you're right confocal microscopic images should be in focus by > definition. It's for sure that our microscope needs to undergo > improvements... > Anyhow, due to a large screening mode of this microscope, I have to cope > with out of focus images (unfortunately, I'm not that deep into physics > to be able to explain the reason for this focus problems). > I don't have stacks. > > Ciao, > Antje > > > > > Karsten Rodenacker schrieb: >> You are tackling an interesting question. To be in focus is always >> meant for an object, adjust focus is meant for a whole image. Hence >> the problem with focus is the observer, a cell of certain thickness >> might be in focus at the cell border or at the cell center. >> >> However, still there is a more important question. Images from >> confocal microscopes are by definition in focus of course related to >> the 'pin-hole' size. If you call such images in 'bad focus' maybe >> there is some help in microscope adjustment necessary? >> If you have stacks of images, images taken at different focal >> adjustments there is a program to merge such stacks into one image >> (e.g. plugin Exteded depth of field). However, my first remark is >> still valid, meaning the outcome of such a method is possibly far away >> from expectation. >> Regards >> Karsten >> >> Am 06.03.2007 um 09:17 schrieb Antje: >> >>> Hello, >>> >>> sorry, the question is not that strongly related with ImageJ but >>> maybe there is still someone who can help? >>> >>> I have a huge set of microscopic images from a confocal microscope >>> (cells/nuclei + marker) and I'd like to automatically filter images >>> which have a bad focus. First, I was thinking about some gradient >>> values to judge but they are so dependent of the content (number of >>> cells... , presence of the marker). Is there any reliable method >>> known to measure and compare the focus of images??? Or does anybody >>> know papers dealing with this issue??? >>> (If there is something available in ImageJ it would be fine but I >>> could not find anything...) >>> >>> Antje >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> > > > > ___________________________________________________________ Telefonate > ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
Hi Jan,
if I take a Sobel filter, it'll be dependent of the amount of objects within the image. Mean, Median or sum of the gradient values over the whole image will be influenced by the number of objects (density), I guess. Further, object detection before, might be dependent on the focus so that you could end up with values you cannot compare. Is there any other way to use the gradient information? Ciao, Antje Jan Eglinger schrieb: > Dear Antje, > > I'd suggest to use an edge detection filter (e.g. Sobel filter "Find > Edges") and to take the brightness of the result as a score. > Images with sharp edges should produce a high score that permits you to > sort your images. > > I don't know if this will work with your set of images, but it should be > a quick way to get some idea of the "focus" quality (provided your > objects are mainly in a single plane of focus). > > Best, > Jan > > -------- Original Message -------- > From: Antje <[hidden email]> > To: [hidden email] > Subject: Re:Focus measurement (more general image analysis question) > Date: 06.03.2007 10:04 > >> Hi Karsten, >> >> you're right confocal microscopic images should be in focus by >> definition. It's for sure that our microscope needs to undergo >> improvements... >> Anyhow, due to a large screening mode of this microscope, I have to >> cope with out of focus images (unfortunately, I'm not that deep into >> physics to be able to explain the reason for this focus problems). >> I don't have stacks. >> >> Ciao, >> Antje >> >> >> >> >> Karsten Rodenacker schrieb: >>> You are tackling an interesting question. To be in focus is always >>> meant for an object, adjust focus is meant for a whole image. Hence >>> the problem with focus is the observer, a cell of certain thickness >>> might be in focus at the cell border or at the cell center. >>> >>> However, still there is a more important question. Images from >>> confocal microscopes are by definition in focus of course related to >>> the 'pin-hole' size. If you call such images in 'bad focus' maybe >>> there is some help in microscope adjustment necessary? >>> If you have stacks of images, images taken at different focal >>> adjustments there is a program to merge such stacks into one image >>> (e.g. plugin Exteded depth of field). However, my first remark is >>> still valid, meaning the outcome of such a method is possibly far >>> away from expectation. >>> Regards >>> Karsten >>> >>> Am 06.03.2007 um 09:17 schrieb Antje: >>> >>>> Hello, >>>> >>>> sorry, the question is not that strongly related with ImageJ but >>>> maybe there is still someone who can help? >>>> >>>> I have a huge set of microscopic images from a confocal microscope >>>> (cells/nuclei + marker) and I'd like to automatically filter images >>>> which have a bad focus. First, I was thinking about some gradient >>>> values to judge but they are so dependent of the content (number of >>>> cells... , presence of the marker). Is there any reliable method >>>> known to measure and compare the focus of images??? Or does anybody >>>> know papers dealing with this issue??? >>>> (If there is something available in ImageJ it would be fine but I >>>> could not find anything...) >>>> >>>> Antje >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >>> >> >> >> ___________________________________________________________ >> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de > ___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de |
|
|
Hi Antje,
a quite difficult discussion track. Perhaps you should explain what you would like to do. I tried to say that there is no focus per se. You should describe for what which focus is necessary. If you like to estimate maker density per cell number or cell volume the focus might be not so important. Since number estimation is expensive (time consuming) possibly the density per volume is sufficient. I have had very often discussions about necessity! Unbiased Stereology from Howard and Reed might help. Regards Karsten Am 06.03.2007 um 10:51 schrieb Antje: > Hi Jan, > > if I take a Sobel filter, it'll be dependent of the amount of > objects within the image. Mean, Median or sum of the gradient > values over the whole image will be influenced by the number of > objects (density), I guess. Further, object detection before, might > be dependent on the focus so that you could end up with values you > cannot compare. > Is there any other way to use the gradient information? > > Ciao, > Antje > > > > Jan Eglinger schrieb: >> Dear Antje, >> I'd suggest to use an edge detection filter (e.g. Sobel filter >> "Find Edges") and to take the brightness of the result as a score. >> Images with sharp edges should produce a high score that permits >> you to sort your images. >> I don't know if this will work with your set of images, but it >> should be a quick way to get some idea of the "focus" quality >> (provided your objects are mainly in a single plane of focus). >> Best, >> Jan >> -------- Original Message -------- >> From: Antje <[hidden email]> >> To: [hidden email] >> Subject: Re:Focus measurement (more general image analysis question) >> Date: 06.03.2007 10:04 >>> Hi Karsten, >>> >>> you're right confocal microscopic images should be in focus by >>> definition. It's for sure that our microscope needs to undergo >>> improvements... >>> Anyhow, due to a large screening mode of this microscope, I have >>> to cope with out of focus images (unfortunately, I'm not that >>> deep into physics to be able to explain the reason for this focus >>> problems). >>> I don't have stacks. >>> >>> Ciao, >>> Antje >>> >>> >>> >>> >>> Karsten Rodenacker schrieb: >>>> You are tackling an interesting question. To be in focus is >>>> always meant for an object, adjust focus is meant for a whole >>>> image. Hence the problem with focus is the observer, a cell of >>>> certain thickness might be in focus at the cell border or at the >>>> cell center. >>>> >>>> However, still there is a more important question. Images from >>>> confocal microscopes are by definition in focus of course >>>> related to the 'pin-hole' size. If you call such images in 'bad >>>> focus' maybe there is some help in microscope adjustment necessary? >>>> If you have stacks of images, images taken at different focal >>>> adjustments there is a program to merge such stacks into one >>>> image (e.g. plugin Exteded depth of field). However, my first >>>> remark is still valid, meaning the outcome of such a method is >>>> possibly far away from expectation. >>>> Regards >>>> Karsten >>>> >>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>> >>>>> Hello, >>>>> >>>>> sorry, the question is not that strongly related with ImageJ >>>>> but maybe there is still someone who can help? >>>>> >>>>> I have a huge set of microscopic images from a confocal >>>>> microscope (cells/nuclei + marker) and I'd like to >>>>> automatically filter images which have a bad focus. First, I >>>>> was thinking about some gradient values to judge but they are >>>>> so dependent of the content (number of cells... , presence of >>>>> the marker). Is there any reliable method known to measure and >>>>> compare the focus of images??? Or does anybody know papers >>>>> dealing with this issue??? >>>>> (If there is something available in ImageJ it would be fine but >>>>> I could not find anything...) >>>>> >>>>> Antje >>>>> >>>>> >>>>> ___________________________________________________________ >>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>> messenger.yahoo.de >>>> >>> >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>> messenger.yahoo.de > > > > > > > ___________________________________________________________ Der > frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! > Mail: http://mail.yahoo.de |
|
|
Hi Karsten,
okay, I try to explain what I would like to do. Let's assume I have 2 images (gray-values) of different fluorochromes of the same objects. On is representing the nucleus and the cell, the other represents a dotty structure within the cell (vesicles). Now, I would like to detect the cell, the nucleus and the vesicles, so that I can measure them (e.g. size of the vesicles). But I would like to exclude that I get biased data from objects which are bad focused. A blurred vesicle then is much larger, than a sharp one, though they might have the same size in reality... My aim is to do high content analysis in an automated way (maybe also extract further parameters of the vesicles and to feed some kind of classifier with the data). Ciao, Antje Karsten Rodenacker schrieb: > Hi Antje, > a quite difficult discussion track. Perhaps you should explain what you > would like to do. I tried to say that there is no focus per se. You > should describe for what which focus is necessary. > If you like to estimate maker density per cell number or cell volume the > focus might be not so important. Since number estimation is expensive > (time consuming) possibly the density per volume is sufficient. I have > had very often discussions about necessity! Unbiased Stereology from > Howard and Reed might help. > Regards > Karsten > > Am 06.03.2007 um 10:51 schrieb Antje: > >> Hi Jan, >> >> if I take a Sobel filter, it'll be dependent of the amount of objects >> within the image. Mean, Median or sum of the gradient values over the >> whole image will be influenced by the number of objects (density), I >> guess. Further, object detection before, might be dependent on the >> focus so that you could end up with values you cannot compare. >> Is there any other way to use the gradient information? >> >> Ciao, >> Antje >> >> >> >> Jan Eglinger schrieb: >>> Dear Antje, >>> I'd suggest to use an edge detection filter (e.g. Sobel filter "Find >>> Edges") and to take the brightness of the result as a score. >>> Images with sharp edges should produce a high score that permits you >>> to sort your images. >>> I don't know if this will work with your set of images, but it should >>> be a quick way to get some idea of the "focus" quality (provided your >>> objects are mainly in a single plane of focus). >>> Best, >>> Jan >>> -------- Original Message -------- >>> From: Antje <[hidden email]> >>> To: [hidden email] >>> Subject: Re:Focus measurement (more general image analysis question) >>> Date: 06.03.2007 10:04 >>>> Hi Karsten, >>>> >>>> you're right confocal microscopic images should be in focus by >>>> definition. It's for sure that our microscope needs to undergo >>>> improvements... >>>> Anyhow, due to a large screening mode of this microscope, I have to >>>> cope with out of focus images (unfortunately, I'm not that deep into >>>> physics to be able to explain the reason for this focus problems). >>>> I don't have stacks. >>>> >>>> Ciao, >>>> Antje >>>> >>>> >>>> >>>> >>>> Karsten Rodenacker schrieb: >>>>> You are tackling an interesting question. To be in focus is always >>>>> meant for an object, adjust focus is meant for a whole image. Hence >>>>> the problem with focus is the observer, a cell of certain thickness >>>>> might be in focus at the cell border or at the cell center. >>>>> >>>>> However, still there is a more important question. Images from >>>>> confocal microscopes are by definition in focus of course related >>>>> to the 'pin-hole' size. If you call such images in 'bad focus' >>>>> maybe there is some help in microscope adjustment necessary? >>>>> If you have stacks of images, images taken at different focal >>>>> adjustments there is a program to merge such stacks into one image >>>>> (e.g. plugin Exteded depth of field). However, my first remark is >>>>> still valid, meaning the outcome of such a method is possibly far >>>>> away from expectation. >>>>> Regards >>>>> Karsten >>>>> >>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>> >>>>>> Hello, >>>>>> >>>>>> sorry, the question is not that strongly related with ImageJ but >>>>>> maybe there is still someone who can help? >>>>>> >>>>>> I have a huge set of microscopic images from a confocal microscope >>>>>> (cells/nuclei + marker) and I'd like to automatically filter >>>>>> images which have a bad focus. First, I was thinking about some >>>>>> gradient values to judge but they are so dependent of the content >>>>>> (number of cells... , presence of the marker). Is there any >>>>>> reliable method known to measure and compare the focus of >>>>>> images??? Or does anybody know papers dealing with this issue??? >>>>>> (If there is something available in ImageJ it would be fine but I >>>>>> could not find anything...) >>>>>> >>>>>> Antje >>>>>> >>>>>> ___________________________________________________________ >>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>> http://messenger.yahoo.de >>>>> >>>> >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> >> >> >> >> >> >> ___________________________________________________________ Der frühe >> Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: >> http://mail.yahoo.de > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
Ok, now we come to some progress, hopefully.
You have the detected cell/nuclei, than -apply the already recommended Sobel or any other edge detector to the vesicle fluorescence image, -Measure SD and mean Gray Value (Redirect To the filtered image) (don't ask me how that is done, but I know it is possible) -find the threshold(s) for which YOU consider the vesicle as bad focussed -throw away the objects falling under the criteria typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. Regards Karsten Am 06.03.2007 um 11:42 schrieb Antje: > Hi Karsten, > > okay, I try to explain what I would like to do. > Let's assume I have 2 images (gray-values) of different > fluorochromes of the same objects. On is representing the nucleus > and the cell, the other represents a dotty structure within the > cell (vesicles). > Now, I would like to detect the cell, the nucleus and the vesicles, > so that I can measure them (e.g. size of the vesicles). > But I would like to exclude that I get biased data from objects > which are bad focused. A blurred vesicle then is much larger, than > a sharp one, though they might have the same size in reality... > My aim is to do high content analysis in an automated way (maybe > also extract further parameters of the vesicles and to feed some > kind of classifier with the data). > > Ciao, > Antje > > > > Karsten Rodenacker schrieb: >> Hi Antje, >> a quite difficult discussion track. Perhaps you should explain >> what you would like to do. I tried to say that there is no focus >> per se. You should describe for what which focus is necessary. >> If you like to estimate maker density per cell number or cell >> volume the focus might be not so important. Since number >> estimation is expensive (time consuming) possibly the density per >> volume is sufficient. I have had very often discussions about >> necessity! Unbiased Stereology from Howard and Reed might help. >> Regards >> Karsten >> Am 06.03.2007 um 10:51 schrieb Antje: >>> Hi Jan, >>> >>> if I take a Sobel filter, it'll be dependent of the amount of >>> objects within the image. Mean, Median or sum of the gradient >>> values over the whole image will be influenced by the number of >>> objects (density), I guess. Further, object detection before, >>> might be dependent on the focus so that you could end up with >>> values you cannot compare. >>> Is there any other way to use the gradient information? >>> >>> Ciao, >>> Antje >>> >>> >>> >>> Jan Eglinger schrieb: >>>> Dear Antje, >>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>> "Find Edges") and to take the brightness of the result as a score. >>>> Images with sharp edges should produce a high score that permits >>>> you to sort your images. >>>> I don't know if this will work with your set of images, but it >>>> should be a quick way to get some idea of the "focus" quality >>>> (provided your objects are mainly in a single plane of focus). >>>> Best, >>>> Jan >>>> -------- Original Message -------- >>>> From: Antje <[hidden email]> >>>> To: [hidden email] >>>> Subject: Re:Focus measurement (more general image analysis >>>> question) >>>> Date: 06.03.2007 10:04 >>>>> Hi Karsten, >>>>> >>>>> you're right confocal microscopic images should be in focus by >>>>> definition. It's for sure that our microscope needs to undergo >>>>> improvements... >>>>> Anyhow, due to a large screening mode of this microscope, I >>>>> have to cope with out of focus images (unfortunately, I'm not >>>>> that deep into physics to be able to explain the reason for >>>>> this focus problems). >>>>> I don't have stacks. >>>>> >>>>> Ciao, >>>>> Antje >>>>> >>>>> >>>>> >>>>> >>>>> Karsten Rodenacker schrieb: >>>>>> You are tackling an interesting question. To be in focus is >>>>>> always meant for an object, adjust focus is meant for a whole >>>>>> image. Hence the problem with focus is the observer, a cell of >>>>>> certain thickness might be in focus at the cell border or at >>>>>> the cell center. >>>>>> >>>>>> However, still there is a more important question. Images from >>>>>> confocal microscopes are by definition in focus of course >>>>>> related to the 'pin-hole' size. If you call such images in >>>>>> 'bad focus' maybe there is some help in microscope adjustment >>>>>> necessary? >>>>>> If you have stacks of images, images taken at different focal >>>>>> adjustments there is a program to merge such stacks into one >>>>>> image (e.g. plugin Exteded depth of field). However, my first >>>>>> remark is still valid, meaning the outcome of such a method is >>>>>> possibly far away from expectation. >>>>>> Regards >>>>>> Karsten >>>>>> >>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>> >>>>>>> Hello, >>>>>>> >>>>>>> sorry, the question is not that strongly related with ImageJ >>>>>>> but maybe there is still someone who can help? >>>>>>> >>>>>>> I have a huge set of microscopic images from a confocal >>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>> automatically filter images which have a bad focus. First, I >>>>>>> was thinking about some gradient values to judge but they are >>>>>>> so dependent of the content (number of cells... , presence of >>>>>>> the marker). Is there any reliable method known to measure >>>>>>> and compare the focus of images??? Or does anybody know >>>>>>> papers dealing with this issue??? >>>>>>> (If there is something available in ImageJ it would be fine >>>>>>> but I could not find anything...) >>>>>>> >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>> messenger.yahoo.de >>>>>> >>>>> >>>>> >>>>> >>>>> ___________________________________________________________ >>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>> messenger.yahoo.de >>> >>> >>> >>> >>> >>> ___________________________________________________________ Der >>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>> Mail: http://mail.yahoo.de > > > > ___________________________________________________________ > Telefonate ohne weitere Kosten vom PC zum PC: http:// > messenger.yahoo.de |
|
|
Hi Karsten,
I don't know whether I understood... (maybe I made a diffuse description)... I don't want to filter objects within one image, I would like to judge whether the majority of objects is in focus or not. In general, I get images which are blurred at all or which have a nice focus over the whole area. I just would like to analyze the images which have well focused objects. And I have to automate this. That means, I'd like something like an "out-of-focus" filter over a set of several thousand images, which delivers only the images with a good focus... Antje Karsten Rodenacker schrieb: > Ok, now we come to some progress, hopefully. > > You have the detected cell/nuclei, than > -apply the already recommended Sobel or any other edge detector to the > vesicle fluorescence image, > -Measure SD and mean Gray Value (Redirect To the filtered image) (don't > ask me how that is done, but I know it is possible) > -find the threshold(s) for which YOU consider the vesicle as bad focussed > -throw away the objects falling under the criteria > typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. > > Regards > Karsten > > Am 06.03.2007 um 11:42 schrieb Antje: > >> Hi Karsten, >> >> okay, I try to explain what I would like to do. >> Let's assume I have 2 images (gray-values) of different fluorochromes >> of the same objects. On is representing the nucleus and the cell, the >> other represents a dotty structure within the cell (vesicles). >> Now, I would like to detect the cell, the nucleus and the vesicles, so >> that I can measure them (e.g. size of the vesicles). >> But I would like to exclude that I get biased data from objects which >> are bad focused. A blurred vesicle then is much larger, than a sharp >> one, though they might have the same size in reality... >> My aim is to do high content analysis in an automated way (maybe also >> extract further parameters of the vesicles and to feed some kind of >> classifier with the data). >> >> Ciao, >> Antje >> >> >> >> Karsten Rodenacker schrieb: >>> Hi Antje, >>> a quite difficult discussion track. Perhaps you should explain what >>> you would like to do. I tried to say that there is no focus per se. >>> You should describe for what which focus is necessary. >>> If you like to estimate maker density per cell number or cell volume >>> the focus might be not so important. Since number estimation is >>> expensive (time consuming) possibly the density per volume is >>> sufficient. I have had very often discussions about necessity! >>> Unbiased Stereology from Howard and Reed might help. >>> Regards >>> Karsten >>> Am 06.03.2007 um 10:51 schrieb Antje: >>>> Hi Jan, >>>> >>>> if I take a Sobel filter, it'll be dependent of the amount of >>>> objects within the image. Mean, Median or sum of the gradient values >>>> over the whole image will be influenced by the number of objects >>>> (density), I guess. Further, object detection before, might be >>>> dependent on the focus so that you could end up with values you >>>> cannot compare. >>>> Is there any other way to use the gradient information? >>>> >>>> Ciao, >>>> Antje >>>> >>>> >>>> >>>> Jan Eglinger schrieb: >>>>> Dear Antje, >>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>> "Find Edges") and to take the brightness of the result as a score. >>>>> Images with sharp edges should produce a high score that permits >>>>> you to sort your images. >>>>> I don't know if this will work with your set of images, but it >>>>> should be a quick way to get some idea of the "focus" quality >>>>> (provided your objects are mainly in a single plane of focus). >>>>> Best, >>>>> Jan >>>>> -------- Original Message -------- >>>>> From: Antje <[hidden email]> >>>>> To: [hidden email] >>>>> Subject: Re:Focus measurement (more general image analysis question) >>>>> Date: 06.03.2007 10:04 >>>>>> Hi Karsten, >>>>>> >>>>>> you're right confocal microscopic images should be in focus by >>>>>> definition. It's for sure that our microscope needs to undergo >>>>>> improvements... >>>>>> Anyhow, due to a large screening mode of this microscope, I have >>>>>> to cope with out of focus images (unfortunately, I'm not that deep >>>>>> into physics to be able to explain the reason for this focus >>>>>> problems). >>>>>> I don't have stacks. >>>>>> >>>>>> Ciao, >>>>>> Antje >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> Karsten Rodenacker schrieb: >>>>>>> You are tackling an interesting question. To be in focus is >>>>>>> always meant for an object, adjust focus is meant for a whole >>>>>>> image. Hence the problem with focus is the observer, a cell of >>>>>>> certain thickness might be in focus at the cell border or at the >>>>>>> cell center. >>>>>>> >>>>>>> However, still there is a more important question. Images from >>>>>>> confocal microscopes are by definition in focus of course related >>>>>>> to the 'pin-hole' size. If you call such images in 'bad focus' >>>>>>> maybe there is some help in microscope adjustment necessary? >>>>>>> If you have stacks of images, images taken at different focal >>>>>>> adjustments there is a program to merge such stacks into one >>>>>>> image (e.g. plugin Exteded depth of field). However, my first >>>>>>> remark is still valid, meaning the outcome of such a method is >>>>>>> possibly far away from expectation. >>>>>>> Regards >>>>>>> Karsten >>>>>>> >>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>> >>>>>>>> Hello, >>>>>>>> >>>>>>>> sorry, the question is not that strongly related with ImageJ but >>>>>>>> maybe there is still someone who can help? >>>>>>>> >>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>> microscope (cells/nuclei + marker) and I'd like to automatically >>>>>>>> filter images which have a bad focus. First, I was thinking >>>>>>>> about some gradient values to judge but they are so dependent of >>>>>>>> the content (number of cells... , presence of the marker). Is >>>>>>>> there any reliable method known to measure and compare the focus >>>>>>>> of images??? Or does anybody know papers dealing with this issue??? >>>>>>>> (If there is something available in ImageJ it would be fine but >>>>>>>> I could not find anything...) >>>>>>>> >>>>>>>> Antje >>>>>>>> >>>>>>>> >>>>>>>> ___________________________________________________________ >>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>> http://messenger.yahoo.de >>>>>>> >>>>>> >>>>>> >>>>>> ___________________________________________________________ >>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>> http://messenger.yahoo.de >>>> >>>> >>>> >>>> >>>> >>>> ___________________________________________________________ Der >>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: >>>> http://mail.yahoo.de >> >> >> >> ___________________________________________________________ Telefonate >> ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
Tscha, maybe I am thinking to complicated...
Possibly you need only a smoothing filter and look than for the difference (again SD and mean). A large SD shows that the original and the smoothed one are different, hence in focus, a small one show the opposite. From that it should be possible to deduce the rule for decision. For smoothing you can apply Smooth or FFT/inverse FFT or Graymorphology Open or Close or Median or ... All is dependent on the blurring. If it is simple, symmetrical the linear operators are best suited. If it is not simple the non-linear morphological/rank order ops are good. They are sensitive either for bright sparks (open) or for dark sparks (close). Regards Karsten Am 06.03.2007 um 14:41 schrieb Antje: > Hi Karsten, > > I don't know whether I understood... (maybe I made a diffuse > description)... I don't want to filter objects within one image, I > would like to judge whether the majority of objects is in focus or > not. In general, I get images which are blurred at all or which > have a nice focus over the whole area. I just would like to analyze > the images which have well focused objects. And I have to automate > this. That means, I'd like something like an "out-of-focus" filter > over a set of several thousand images, which delivers only the > images with a good focus... > > Antje > > > > Karsten Rodenacker schrieb: >> Ok, now we come to some progress, hopefully. >> You have the detected cell/nuclei, than >> -apply the already recommended Sobel or any other edge detector to >> the vesicle fluorescence image, >> -Measure SD and mean Gray Value (Redirect To the filtered image) >> (don't ask me how that is done, but I know it is possible) >> -find the threshold(s) for which YOU consider the vesicle as bad >> focussed >> -throw away the objects falling under the criteria >> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >> Regards >> Karsten >> Am 06.03.2007 um 11:42 schrieb Antje: >>> Hi Karsten, >>> >>> okay, I try to explain what I would like to do. >>> Let's assume I have 2 images (gray-values) of different >>> fluorochromes of the same objects. On is representing the nucleus >>> and the cell, the other represents a dotty structure within the >>> cell (vesicles). >>> Now, I would like to detect the cell, the nucleus and the >>> vesicles, so that I can measure them (e.g. size of the vesicles). >>> But I would like to exclude that I get biased data from objects >>> which are bad focused. A blurred vesicle then is much larger, >>> than a sharp one, though they might have the same size in reality... >>> My aim is to do high content analysis in an automated way (maybe >>> also extract further parameters of the vesicles and to feed some >>> kind of classifier with the data). >>> >>> Ciao, >>> Antje >>> >>> >>> >>> Karsten Rodenacker schrieb: >>>> Hi Antje, >>>> a quite difficult discussion track. Perhaps you should explain >>>> what you would like to do. I tried to say that there is no focus >>>> per se. You should describe for what which focus is necessary. >>>> If you like to estimate maker density per cell number or cell >>>> volume the focus might be not so important. Since number >>>> estimation is expensive (time consuming) possibly the density >>>> per volume is sufficient. I have had very often discussions >>>> about necessity! Unbiased Stereology from Howard and Reed might >>>> help. >>>> Regards >>>> Karsten >>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>> Hi Jan, >>>>> >>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>> objects within the image. Mean, Median or sum of the gradient >>>>> values over the whole image will be influenced by the number of >>>>> objects (density), I guess. Further, object detection before, >>>>> might be dependent on the focus so that you could end up with >>>>> values you cannot compare. >>>>> Is there any other way to use the gradient information? >>>>> >>>>> Ciao, >>>>> Antje >>>>> >>>>> >>>>> >>>>> Jan Eglinger schrieb: >>>>>> Dear Antje, >>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>> "Find Edges") and to take the brightness of the result as a >>>>>> score. >>>>>> Images with sharp edges should produce a high score that >>>>>> permits you to sort your images. >>>>>> I don't know if this will work with your set of images, but it >>>>>> should be a quick way to get some idea of the "focus" quality >>>>>> (provided your objects are mainly in a single plane of focus). >>>>>> Best, >>>>>> Jan >>>>>> -------- Original Message -------- >>>>>> From: Antje <[hidden email]> >>>>>> To: [hidden email] >>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>> question) >>>>>> Date: 06.03.2007 10:04 >>>>>>> Hi Karsten, >>>>>>> >>>>>>> you're right confocal microscopic images should be in focus >>>>>>> by definition. It's for sure that our microscope needs to >>>>>>> undergo improvements... >>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>> have to cope with out of focus images (unfortunately, I'm not >>>>>>> that deep into physics to be able to explain the reason for >>>>>>> this focus problems). >>>>>>> I don't have stacks. >>>>>>> >>>>>>> Ciao, >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> Karsten Rodenacker schrieb: >>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>> always meant for an object, adjust focus is meant for a >>>>>>>> whole image. Hence the problem with focus is the observer, a >>>>>>>> cell of certain thickness might be in focus at the cell >>>>>>>> border or at the cell center. >>>>>>>> >>>>>>>> However, still there is a more important question. Images >>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>> images in 'bad focus' maybe there is some help in microscope >>>>>>>> adjustment necessary? >>>>>>>> If you have stacks of images, images taken at different >>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>> However, my first remark is still valid, meaning the outcome >>>>>>>> of such a method is possibly far away from expectation. >>>>>>>> Regards >>>>>>>> Karsten >>>>>>>> >>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>> >>>>>>>>> Hello, >>>>>>>>> >>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>> >>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>> automatically filter images which have a bad focus. First, >>>>>>>>> I was thinking about some gradient values to judge but they >>>>>>>>> are so dependent of the content (number of cells... , >>>>>>>>> presence of the marker). Is there any reliable method known >>>>>>>>> to measure and compare the focus of images??? Or does >>>>>>>>> anybody know papers dealing with this issue??? >>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>> but I could not find anything...) >>>>>>>>> >>>>>>>>> Antje >>>>>>>>> >>>>>>>>> >>>>>>>>> ___________________________________________________________ >>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>> messenger.yahoo.de >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>> messenger.yahoo.de >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> ___________________________________________________________ Der >>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>> Mail: http://mail.yahoo.de >>> >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>> messenger.yahoo.de > > > > ___________________________________________________________ > Telefonate ohne weitere Kosten vom PC zum PC: http:// > messenger.yahoo.de |
Re: Focus measurement (more general image analysis question)
|
|
Hi, I can think of another solution. Perhaps write a plug-in that displays a small GUI that can sequentially open / display the files in a folder. Then the operator uses a "Save/Open-Next" button on the GUI to save the ones in-focus to a separate folder or a "Open-Next" button to just move onto the next image. This may not be as elegant as filters (makes the operator the filter which also may introduce bias) but if there is a reasonably finite number of images might be the fastest and maybe even more reliable.
--- Mike -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Karsten Rodenacker Sent: Tuesday, March 06, 2007 9:03 AM To: [hidden email] Subject: Re: Focus measurement (more general image analysis question) Tscha, maybe I am thinking to complicated... Possibly you need only a smoothing filter and look than for the difference (again SD and mean). A large SD shows that the original and the smoothed one are different, hence in focus, a small one show the opposite. From that it should be possible to deduce the rule for decision. For smoothing you can apply Smooth or FFT/inverse FFT or Graymorphology Open or Close or Median or ... All is dependent on the blurring. If it is simple, symmetrical the linear operators are best suited. If it is not simple the non-linear morphological/rank order ops are good. They are sensitive either for bright sparks (open) or for dark sparks (close). Regards Karsten Am 06.03.2007 um 14:41 schrieb Antje: > Hi Karsten, > > I don't know whether I understood... (maybe I made a diffuse > description)... I don't want to filter objects within one image, I > would like to judge whether the majority of objects is in focus or > not. In general, I get images which are blurred at all or which have a > nice focus over the whole area. I just would like to analyze the > images which have well focused objects. And I have to automate this. > That means, I'd like something like an "out-of-focus" filter over a > set of several thousand images, which delivers only the images with a > good focus... > > Antje > > > > Karsten Rodenacker schrieb: >> Ok, now we come to some progress, hopefully. >> You have the detected cell/nuclei, than >> -apply the already recommended Sobel or any other edge detector to >> the vesicle fluorescence image, >> -Measure SD and mean Gray Value (Redirect To the filtered image) >> (don't ask me how that is done, but I know it is possible) >> -find the threshold(s) for which YOU consider the vesicle as bad >> focussed >> -throw away the objects falling under the criteria >> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >> Regards >> Karsten >> Am 06.03.2007 um 11:42 schrieb Antje: >>> Hi Karsten, >>> >>> okay, I try to explain what I would like to do. >>> Let's assume I have 2 images (gray-values) of different >>> fluorochromes of the same objects. On is representing the nucleus >>> and the cell, the other represents a dotty structure within the >>> cell (vesicles). >>> Now, I would like to detect the cell, the nucleus and the >>> vesicles, so that I can measure them (e.g. size of the vesicles). >>> But I would like to exclude that I get biased data from objects >>> which are bad focused. A blurred vesicle then is much larger, >>> than a sharp one, though they might have the same size in reality... >>> My aim is to do high content analysis in an automated way (maybe >>> also extract further parameters of the vesicles and to feed some >>> kind of classifier with the data). >>> >>> Ciao, >>> Antje >>> >>> >>> >>> Karsten Rodenacker schrieb: >>>> Hi Antje, >>>> a quite difficult discussion track. Perhaps you should explain >>>> what you would like to do. I tried to say that there is no focus >>>> per se. You should describe for what which focus is necessary. >>>> If you like to estimate maker density per cell number or cell >>>> volume the focus might be not so important. Since number >>>> estimation is expensive (time consuming) possibly the density >>>> per volume is sufficient. I have had very often discussions >>>> about necessity! Unbiased Stereology from Howard and Reed might >>>> help. >>>> Regards >>>> Karsten >>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>> Hi Jan, >>>>> >>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>> objects within the image. Mean, Median or sum of the gradient >>>>> values over the whole image will be influenced by the number of >>>>> objects (density), I guess. Further, object detection before, >>>>> might be dependent on the focus so that you could end up with >>>>> values you cannot compare. >>>>> Is there any other way to use the gradient information? >>>>> >>>>> Ciao, >>>>> Antje >>>>> >>>>> >>>>> >>>>> Jan Eglinger schrieb: >>>>>> Dear Antje, >>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>> "Find Edges") and to take the brightness of the result as a >>>>>> score. >>>>>> Images with sharp edges should produce a high score that >>>>>> permits you to sort your images. >>>>>> I don't know if this will work with your set of images, but it >>>>>> should be a quick way to get some idea of the "focus" quality >>>>>> (provided your objects are mainly in a single plane of focus). >>>>>> Best, >>>>>> Jan >>>>>> -------- Original Message -------- >>>>>> From: Antje <[hidden email]> >>>>>> To: [hidden email] >>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>> question) >>>>>> Date: 06.03.2007 10:04 >>>>>>> Hi Karsten, >>>>>>> >>>>>>> you're right confocal microscopic images should be in focus >>>>>>> by definition. It's for sure that our microscope needs to >>>>>>> undergo improvements... >>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>> have to cope with out of focus images (unfortunately, I'm not >>>>>>> that deep into physics to be able to explain the reason for >>>>>>> this focus problems). >>>>>>> I don't have stacks. >>>>>>> >>>>>>> Ciao, >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> Karsten Rodenacker schrieb: >>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>> always meant for an object, adjust focus is meant for a >>>>>>>> whole image. Hence the problem with focus is the observer, a >>>>>>>> cell of certain thickness might be in focus at the cell >>>>>>>> border or at the cell center. >>>>>>>> >>>>>>>> However, still there is a more important question. Images >>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>> images in 'bad focus' maybe there is some help in microscope >>>>>>>> adjustment necessary? >>>>>>>> If you have stacks of images, images taken at different >>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>> However, my first remark is still valid, meaning the outcome >>>>>>>> of such a method is possibly far away from expectation. >>>>>>>> Regards >>>>>>>> Karsten >>>>>>>> >>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>> >>>>>>>>> Hello, >>>>>>>>> >>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>> >>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>> automatically filter images which have a bad focus. First, >>>>>>>>> I was thinking about some gradient values to judge but they >>>>>>>>> are so dependent of the content (number of cells... , >>>>>>>>> presence of the marker). Is there any reliable method known >>>>>>>>> to measure and compare the focus of images??? Or does >>>>>>>>> anybody know papers dealing with this issue??? >>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>> but I could not find anything...) >>>>>>>>> >>>>>>>>> Antje >>>>>>>>> >>>>>>>>> >>>>>>>>> ___________________________________________________________ >>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>> messenger.yahoo.de >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>> messenger.yahoo.de >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> ___________________________________________________________ Der >>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>> Mail: http://mail.yahoo.de >>> >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>> messenger.yahoo.de > > > > ___________________________________________________________ > Telefonate ohne weitere Kosten vom PC zum PC: http:// > messenger.yahoo.de |
|
|
In reply to this post by karo03
Hi Karsten,
that's a nice idea. But still I have a question. How shall I compare images with different density? Because the standard deviation of an image will be dependent on the density of objects. It may happen, that there is just one cell within one image and it can also happen, that there is no background at all because of the density of cells... Or did I miss something of your idea? Every value I can think of seems to be dependent of the content/background ratio... Ciao, Antje Karsten Rodenacker schrieb: > Tscha, maybe I am thinking to complicated... > > Possibly you need only a smoothing filter and look than for the > difference (again SD and mean). A large SD shows that the original and > the smoothed one are different, hence in focus, a small one show the > opposite. From that it should be possible to deduce the rule for > decision. For smoothing you can apply Smooth or FFT/inverse FFT or > Graymorphology Open or Close or Median or ... All is dependent on the > blurring. If it is simple, symmetrical the linear operators are best > suited. If it is not simple the non-linear morphological/rank order ops > are good. They are sensitive either for bright sparks (open) or for dark > sparks (close). > > Regards > Karsten > > Am 06.03.2007 um 14:41 schrieb Antje: > >> Hi Karsten, >> >> I don't know whether I understood... (maybe I made a diffuse >> description)... I don't want to filter objects within one image, I >> would like to judge whether the majority of objects is in focus or >> not. In general, I get images which are blurred at all or which have a >> nice focus over the whole area. I just would like to analyze the >> images which have well focused objects. And I have to automate this. >> That means, I'd like something like an "out-of-focus" filter over a >> set of several thousand images, which delivers only the images with a >> good focus... >> >> Antje >> >> >> >> Karsten Rodenacker schrieb: >>> Ok, now we come to some progress, hopefully. >>> You have the detected cell/nuclei, than >>> -apply the already recommended Sobel or any other edge detector to >>> the vesicle fluorescence image, >>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>> (don't ask me how that is done, but I know it is possible) >>> -find the threshold(s) for which YOU consider the vesicle as bad >>> focussed >>> -throw away the objects falling under the criteria >>> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >>> Regards >>> Karsten >>> Am 06.03.2007 um 11:42 schrieb Antje: >>>> Hi Karsten, >>>> >>>> okay, I try to explain what I would like to do. >>>> Let's assume I have 2 images (gray-values) of different >>>> fluorochromes of the same objects. On is representing the nucleus >>>> and the cell, the other represents a dotty structure within the cell >>>> (vesicles). >>>> Now, I would like to detect the cell, the nucleus and the vesicles, >>>> so that I can measure them (e.g. size of the vesicles). >>>> But I would like to exclude that I get biased data from objects >>>> which are bad focused. A blurred vesicle then is much larger, than a >>>> sharp one, though they might have the same size in reality... >>>> My aim is to do high content analysis in an automated way (maybe >>>> also extract further parameters of the vesicles and to feed some >>>> kind of classifier with the data). >>>> >>>> Ciao, >>>> Antje >>>> >>>> >>>> >>>> Karsten Rodenacker schrieb: >>>>> Hi Antje, >>>>> a quite difficult discussion track. Perhaps you should explain what >>>>> you would like to do. I tried to say that there is no focus per se. >>>>> You should describe for what which focus is necessary. >>>>> If you like to estimate maker density per cell number or cell >>>>> volume the focus might be not so important. Since number estimation >>>>> is expensive (time consuming) possibly the density per volume is >>>>> sufficient. I have had very often discussions about necessity! >>>>> Unbiased Stereology from Howard and Reed might help. >>>>> Regards >>>>> Karsten >>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>> Hi Jan, >>>>>> >>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>> values over the whole image will be influenced by the number of >>>>>> objects (density), I guess. Further, object detection before, >>>>>> might be dependent on the focus so that you could end up with >>>>>> values you cannot compare. >>>>>> Is there any other way to use the gradient information? >>>>>> >>>>>> Ciao, >>>>>> Antje >>>>>> >>>>>> >>>>>> >>>>>> Jan Eglinger schrieb: >>>>>>> Dear Antje, >>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>>> "Find Edges") and to take the brightness of the result as a score. >>>>>>> Images with sharp edges should produce a high score that permits >>>>>>> you to sort your images. >>>>>>> I don't know if this will work with your set of images, but it >>>>>>> should be a quick way to get some idea of the "focus" quality >>>>>>> (provided your objects are mainly in a single plane of focus). >>>>>>> Best, >>>>>>> Jan >>>>>>> -------- Original Message -------- >>>>>>> From: Antje <[hidden email]> >>>>>>> To: [hidden email] >>>>>>> Subject: Re:Focus measurement (more general image analysis question) >>>>>>> Date: 06.03.2007 10:04 >>>>>>>> Hi Karsten, >>>>>>>> >>>>>>>> you're right confocal microscopic images should be in focus by >>>>>>>> definition. It's for sure that our microscope needs to undergo >>>>>>>> improvements... >>>>>>>> Anyhow, due to a large screening mode of this microscope, I have >>>>>>>> to cope with out of focus images (unfortunately, I'm not that >>>>>>>> deep into physics to be able to explain the reason for this >>>>>>>> focus problems). >>>>>>>> I don't have stacks. >>>>>>>> >>>>>>>> Ciao, >>>>>>>> Antje >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>>> always meant for an object, adjust focus is meant for a whole >>>>>>>>> image. Hence the problem with focus is the observer, a cell of >>>>>>>>> certain thickness might be in focus at the cell border or at >>>>>>>>> the cell center. >>>>>>>>> >>>>>>>>> However, still there is a more important question. Images from >>>>>>>>> confocal microscopes are by definition in focus of course >>>>>>>>> related to the 'pin-hole' size. If you call such images in 'bad >>>>>>>>> focus' maybe there is some help in microscope adjustment >>>>>>>>> necessary? >>>>>>>>> If you have stacks of images, images taken at different focal >>>>>>>>> adjustments there is a program to merge such stacks into one >>>>>>>>> image (e.g. plugin Exteded depth of field). However, my first >>>>>>>>> remark is still valid, meaning the outcome of such a method is >>>>>>>>> possibly far away from expectation. >>>>>>>>> Regards >>>>>>>>> Karsten >>>>>>>>> >>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>> >>>>>>>>>> Hello, >>>>>>>>>> >>>>>>>>>> sorry, the question is not that strongly related with ImageJ >>>>>>>>>> but maybe there is still someone who can help? >>>>>>>>>> >>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>> automatically filter images which have a bad focus. First, I >>>>>>>>>> was thinking about some gradient values to judge but they are >>>>>>>>>> so dependent of the content (number of cells... , presence of >>>>>>>>>> the marker). Is there any reliable method known to measure and >>>>>>>>>> compare the focus of images??? Or does anybody know papers >>>>>>>>>> dealing with this issue??? >>>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>>> but I could not find anything...) >>>>>>>>>> >>>>>>>>>> Antje >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ___________________________________________________________ >>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>> http://messenger.yahoo.de >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ___________________________________________________________ >>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>> http://messenger.yahoo.de >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> ___________________________________________________________ Der >>>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>>> Mail: http://mail.yahoo.de >>>> >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> >> >> >> ___________________________________________________________ Telefonate >> ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
... the difference between the image and the smoothed version, that
is all what is smoothed away, should not be too dependent on intensity! The idea is to apply an operation which is similar to defocussing. The more the original looks like a smoothed one the more the image is out of focus AND the difference will have a smaller deviation. Of course a smooth original without edges and hence without objects cannot be considered as in or out of focus. Have a look also into the other posts! My first idea tended to classify cellular objects, my second one for whole images, if only some objects are out of focus than it has to be decided if this image is acceptable or not. Global (image based) measures will smooth local focal properties! Regards Karsten Am 07.03.2007 um 14:10 schrieb Antje: > Hi Karsten, > > that's a nice idea. But still I have a question. How shall I > compare images with different density? Because the standard > deviation of an image will be dependent on the density of objects. > It may happen, that there is just one cell within one image and it > can also happen, that there is no background at all because of the > density of cells... > Or did I miss something of your idea? > Every value I can think of seems to be dependent of the content/ > background ratio... > > Ciao, > Antje > > > Karsten Rodenacker schrieb: >> Tscha, maybe I am thinking to complicated... >> Possibly you need only a smoothing filter and look than for the >> difference (again SD and mean). A large SD shows that the original >> and the smoothed one are different, hence in focus, a small one >> show the opposite. From that it should be possible to deduce the >> rule for decision. For smoothing you can apply Smooth or FFT/ >> inverse FFT or Graymorphology Open or Close or Median or ... All >> is dependent on the blurring. If it is simple, symmetrical the >> linear operators are best suited. If it is not simple the non- >> linear morphological/rank order ops are good. They are sensitive >> either for bright sparks (open) or for dark sparks (close). >> Regards >> Karsten >> Am 06.03.2007 um 14:41 schrieb Antje: >>> Hi Karsten, >>> >>> I don't know whether I understood... (maybe I made a diffuse >>> description)... I don't want to filter objects within one image, >>> I would like to judge whether the majority of objects is in focus >>> or not. In general, I get images which are blurred at all or >>> which have a nice focus over the whole area. I just would like to >>> analyze the images which have well focused objects. And I have to >>> automate this. That means, I'd like something like an "out-of- >>> focus" filter over a set of several thousand images, which >>> delivers only the images with a good focus... >>> >>> Antje >>> >>> >>> >>> Karsten Rodenacker schrieb: >>>> Ok, now we come to some progress, hopefully. >>>> You have the detected cell/nuclei, than >>>> -apply the already recommended Sobel or any other edge detector >>>> to the vesicle fluorescence image, >>>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>>> (don't ask me how that is done, but I know it is possible) >>>> -find the threshold(s) for which YOU consider the vesicle as bad >>>> focussed >>>> -throw away the objects falling under the criteria >>>> typically SD or SD/Mean can discriminate unfocussed (smoothed) >>>> areas. >>>> Regards >>>> Karsten >>>> Am 06.03.2007 um 11:42 schrieb Antje: >>>>> Hi Karsten, >>>>> >>>>> okay, I try to explain what I would like to do. >>>>> Let's assume I have 2 images (gray-values) of different >>>>> fluorochromes of the same objects. On is representing the >>>>> nucleus and the cell, the other represents a dotty structure >>>>> within the cell (vesicles). >>>>> Now, I would like to detect the cell, the nucleus and the >>>>> vesicles, so that I can measure them (e.g. size of the vesicles). >>>>> But I would like to exclude that I get biased data from objects >>>>> which are bad focused. A blurred vesicle then is much larger, >>>>> than a sharp one, though they might have the same size in >>>>> reality... >>>>> My aim is to do high content analysis in an automated way >>>>> (maybe also extract further parameters of the vesicles and to >>>>> feed some kind of classifier with the data). >>>>> >>>>> Ciao, >>>>> Antje >>>>> >>>>> >>>>> >>>>> Karsten Rodenacker schrieb: >>>>>> Hi Antje, >>>>>> a quite difficult discussion track. Perhaps you should explain >>>>>> what you would like to do. I tried to say that there is no >>>>>> focus per se. You should describe for what which focus is >>>>>> necessary. >>>>>> If you like to estimate maker density per cell number or cell >>>>>> volume the focus might be not so important. Since number >>>>>> estimation is expensive (time consuming) possibly the density >>>>>> per volume is sufficient. I have had very often discussions >>>>>> about necessity! Unbiased Stereology from Howard and Reed >>>>>> might help. >>>>>> Regards >>>>>> Karsten >>>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>>> Hi Jan, >>>>>>> >>>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>>> values over the whole image will be influenced by the number >>>>>>> of objects (density), I guess. Further, object detection >>>>>>> before, might be dependent on the focus so that you could end >>>>>>> up with values you cannot compare. >>>>>>> Is there any other way to use the gradient information? >>>>>>> >>>>>>> Ciao, >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> >>>>>>> Jan Eglinger schrieb: >>>>>>>> Dear Antje, >>>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel >>>>>>>> filter "Find Edges") and to take the brightness of the >>>>>>>> result as a score. >>>>>>>> Images with sharp edges should produce a high score that >>>>>>>> permits you to sort your images. >>>>>>>> I don't know if this will work with your set of images, but >>>>>>>> it should be a quick way to get some idea of the "focus" >>>>>>>> quality (provided your objects are mainly in a single plane >>>>>>>> of focus). >>>>>>>> Best, >>>>>>>> Jan >>>>>>>> -------- Original Message -------- >>>>>>>> From: Antje <[hidden email]> >>>>>>>> To: [hidden email] >>>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>>> question) >>>>>>>> Date: 06.03.2007 10:04 >>>>>>>>> Hi Karsten, >>>>>>>>> >>>>>>>>> you're right confocal microscopic images should be in focus >>>>>>>>> by definition. It's for sure that our microscope needs to >>>>>>>>> undergo improvements... >>>>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>>>> have to cope with out of focus images (unfortunately, I'm >>>>>>>>> not that deep into physics to be able to explain the reason >>>>>>>>> for this focus problems). >>>>>>>>> I don't have stacks. >>>>>>>>> >>>>>>>>> Ciao, >>>>>>>>> Antje >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>>> You are tackling an interesting question. To be in focus >>>>>>>>>> is always meant for an object, adjust focus is meant for a >>>>>>>>>> whole image. Hence the problem with focus is the observer, >>>>>>>>>> a cell of certain thickness might be in focus at the cell >>>>>>>>>> border or at the cell center. >>>>>>>>>> >>>>>>>>>> However, still there is a more important question. Images >>>>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>>>> images in 'bad focus' maybe there is some help in >>>>>>>>>> microscope adjustment necessary? >>>>>>>>>> If you have stacks of images, images taken at different >>>>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>>>> However, my first remark is still valid, meaning the >>>>>>>>>> outcome of such a method is possibly far away from >>>>>>>>>> expectation. >>>>>>>>>> Regards >>>>>>>>>> Karsten >>>>>>>>>> >>>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>>> >>>>>>>>>>> Hello, >>>>>>>>>>> >>>>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>>>> >>>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>>> automatically filter images which have a bad focus. >>>>>>>>>>> First, I was thinking about some gradient values to judge >>>>>>>>>>> but they are so dependent of the content (number of >>>>>>>>>>> cells... , presence of the marker). Is there any reliable >>>>>>>>>>> method known to measure and compare the focus of >>>>>>>>>>> images??? Or does anybody know papers dealing with this >>>>>>>>>>> issue??? >>>>>>>>>>> (If there is something available in ImageJ it would be >>>>>>>>>>> fine but I could not find anything...) >>>>>>>>>>> >>>>>>>>>>> Antje >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>>>> messenger.yahoo.de >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ___________________________________________________________ >>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>> messenger.yahoo.de >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen >>>>>>> Yahoo! Mail: http://mail.yahoo.de >>>>> >>>>> >>>>> >>>>> ___________________________________________________________ >>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>> messenger.yahoo.de >>> >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>> messenger.yahoo.de > > > > ___________________________________________________________ > Telefonate ohne weitere Kosten vom PC zum PC: http:// > messenger.yahoo.de |
|
|
Hi Karsten,
I'm afraid, I still did not get it. Lets assume, I have an image with only one single cell (in focus). Then, of course the difference image shows up quite "high" values for this area. But calculating a mean or a stddev will be biased by the large area having low values (because there is nothing inside). Am I wrong??? It will return different values than an image containing a lot of cells with the same focus), no? which other posts do you mean? ciao, Antje Karsten Rodenacker schrieb: > ... the difference between the image and the smoothed version, that is > all what is smoothed away, should not be too dependent on intensity! The > idea is to apply an operation which is similar to defocussing. The more > the original looks like a smoothed one the more the image is out of > focus AND the difference will have a smaller deviation. Of course a > smooth original without edges and hence without objects cannot be > considered as in or out of focus. > > Have a look also into the other posts! > > My first idea tended to classify cellular objects, my second one for > whole images, if only some objects are out of focus than it has to be > decided if this image is acceptable or not. Global (image based) > measures will smooth local focal properties! > Regards > Karsten > > Am 07.03.2007 um 14:10 schrieb Antje: > >> Hi Karsten, >> >> that's a nice idea. But still I have a question. How shall I compare >> images with different density? Because the standard deviation of an >> image will be dependent on the density of objects. It may happen, that >> there is just one cell within one image and it can also happen, that >> there is no background at all because of the density of cells... >> Or did I miss something of your idea? >> Every value I can think of seems to be dependent of the >> content/background ratio... >> >> Ciao, >> Antje >> >> >> Karsten Rodenacker schrieb: >>> Tscha, maybe I am thinking to complicated... >>> Possibly you need only a smoothing filter and look than for the >>> difference (again SD and mean). A large SD shows that the original >>> and the smoothed one are different, hence in focus, a small one show >>> the opposite. From that it should be possible to deduce the rule for >>> decision. For smoothing you can apply Smooth or FFT/inverse FFT or >>> Graymorphology Open or Close or Median or ... All is dependent on the >>> blurring. If it is simple, symmetrical the linear operators are best >>> suited. If it is not simple the non-linear morphological/rank order >>> ops are good. They are sensitive either for bright sparks (open) or >>> for dark sparks (close). >>> Regards >>> Karsten >>> Am 06.03.2007 um 14:41 schrieb Antje: >>>> Hi Karsten, >>>> >>>> I don't know whether I understood... (maybe I made a diffuse >>>> description)... I don't want to filter objects within one image, I >>>> would like to judge whether the majority of objects is in focus or >>>> not. In general, I get images which are blurred at all or which have >>>> a nice focus over the whole area. I just would like to analyze the >>>> images which have well focused objects. And I have to automate this. >>>> That means, I'd like something like an "out-of-focus" filter over a >>>> set of several thousand images, which delivers only the images with >>>> a good focus... >>>> >>>> Antje >>>> >>>> >>>> >>>> Karsten Rodenacker schrieb: >>>>> Ok, now we come to some progress, hopefully. >>>>> You have the detected cell/nuclei, than >>>>> -apply the already recommended Sobel or any other edge detector to >>>>> the vesicle fluorescence image, >>>>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>>>> (don't ask me how that is done, but I know it is possible) >>>>> -find the threshold(s) for which YOU consider the vesicle as bad >>>>> focussed >>>>> -throw away the objects falling under the criteria >>>>> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >>>>> Regards >>>>> Karsten >>>>> Am 06.03.2007 um 11:42 schrieb Antje: >>>>>> Hi Karsten, >>>>>> >>>>>> okay, I try to explain what I would like to do. >>>>>> Let's assume I have 2 images (gray-values) of different >>>>>> fluorochromes of the same objects. On is representing the nucleus >>>>>> and the cell, the other represents a dotty structure within the >>>>>> cell (vesicles). >>>>>> Now, I would like to detect the cell, the nucleus and the >>>>>> vesicles, so that I can measure them (e.g. size of the vesicles). >>>>>> But I would like to exclude that I get biased data from objects >>>>>> which are bad focused. A blurred vesicle then is much larger, than >>>>>> a sharp one, though they might have the same size in reality... >>>>>> My aim is to do high content analysis in an automated way (maybe >>>>>> also extract further parameters of the vesicles and to feed some >>>>>> kind of classifier with the data). >>>>>> >>>>>> Ciao, >>>>>> Antje >>>>>> >>>>>> >>>>>> >>>>>> Karsten Rodenacker schrieb: >>>>>>> Hi Antje, >>>>>>> a quite difficult discussion track. Perhaps you should explain >>>>>>> what you would like to do. I tried to say that there is no focus >>>>>>> per se. You should describe for what which focus is necessary. >>>>>>> If you like to estimate maker density per cell number or cell >>>>>>> volume the focus might be not so important. Since number >>>>>>> estimation is expensive (time consuming) possibly the density per >>>>>>> volume is sufficient. I have had very often discussions about >>>>>>> necessity! Unbiased Stereology from Howard and Reed might help. >>>>>>> Regards >>>>>>> Karsten >>>>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>>>> Hi Jan, >>>>>>>> >>>>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>>>> values over the whole image will be influenced by the number of >>>>>>>> objects (density), I guess. Further, object detection before, >>>>>>>> might be dependent on the focus so that you could end up with >>>>>>>> values you cannot compare. >>>>>>>> Is there any other way to use the gradient information? >>>>>>>> >>>>>>>> Ciao, >>>>>>>> Antje >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Jan Eglinger schrieb: >>>>>>>>> Dear Antje, >>>>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>>>>> "Find Edges") and to take the brightness of the result as a score. >>>>>>>>> Images with sharp edges should produce a high score that >>>>>>>>> permits you to sort your images. >>>>>>>>> I don't know if this will work with your set of images, but it >>>>>>>>> should be a quick way to get some idea of the "focus" quality >>>>>>>>> (provided your objects are mainly in a single plane of focus). >>>>>>>>> Best, >>>>>>>>> Jan >>>>>>>>> -------- Original Message -------- >>>>>>>>> From: Antje <[hidden email]> >>>>>>>>> To: [hidden email] >>>>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>>>> question) >>>>>>>>> Date: 06.03.2007 10:04 >>>>>>>>>> Hi Karsten, >>>>>>>>>> >>>>>>>>>> you're right confocal microscopic images should be in focus by >>>>>>>>>> definition. It's for sure that our microscope needs to undergo >>>>>>>>>> improvements... >>>>>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>>>>> have to cope with out of focus images (unfortunately, I'm not >>>>>>>>>> that deep into physics to be able to explain the reason for >>>>>>>>>> this focus problems). >>>>>>>>>> I don't have stacks. >>>>>>>>>> >>>>>>>>>> Ciao, >>>>>>>>>> Antje >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>>>>> always meant for an object, adjust focus is meant for a whole >>>>>>>>>>> image. Hence the problem with focus is the observer, a cell >>>>>>>>>>> of certain thickness might be in focus at the cell border or >>>>>>>>>>> at the cell center. >>>>>>>>>>> >>>>>>>>>>> However, still there is a more important question. Images >>>>>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>>>>> images in 'bad focus' maybe there is some help in microscope >>>>>>>>>>> adjustment necessary? >>>>>>>>>>> If you have stacks of images, images taken at different focal >>>>>>>>>>> adjustments there is a program to merge such stacks into one >>>>>>>>>>> image (e.g. plugin Exteded depth of field). However, my first >>>>>>>>>>> remark is still valid, meaning the outcome of such a method >>>>>>>>>>> is possibly far away from expectation. >>>>>>>>>>> Regards >>>>>>>>>>> Karsten >>>>>>>>>>> >>>>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>>>> >>>>>>>>>>>> Hello, >>>>>>>>>>>> >>>>>>>>>>>> sorry, the question is not that strongly related with ImageJ >>>>>>>>>>>> but maybe there is still someone who can help? >>>>>>>>>>>> >>>>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>>>> automatically filter images which have a bad focus. First, I >>>>>>>>>>>> was thinking about some gradient values to judge but they >>>>>>>>>>>> are so dependent of the content (number of cells... , >>>>>>>>>>>> presence of the marker). Is there any reliable method known >>>>>>>>>>>> to measure and compare the focus of images??? Or does >>>>>>>>>>>> anybody know papers dealing with this issue??? >>>>>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>>>>> but I could not find anything...) >>>>>>>>>>>> >>>>>>>>>>>> Antje >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>>>> http://messenger.yahoo.de >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ___________________________________________________________ >>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>> http://messenger.yahoo.de >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ___________________________________________________________ Der >>>>>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>>>>> Mail: http://mail.yahoo.de >>>>>> >>>>>> >>>>>> ___________________________________________________________ >>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>> http://messenger.yahoo.de >>>> >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> >> >> >> ___________________________________________________________ Telefonate >> ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de > ___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de |
|
|
In reply to this post by Sullivan, Michael J (College of Med.)
Hi Michael,
thanks for your answer. I'm afraid this is not an option because I have to handle some thousands of images and I'll have to solve this problem for images which are produced in the future ( more and more and more ...) Ciao, Antje Sullivan, Michael J schrieb: > Hi, I can think of another solution. Perhaps write a plug-in that displays a small GUI that can sequentially open / display the files in a folder. Then the operator uses a "Save/Open-Next" button on the GUI to save the ones in-focus to a separate folder or a "Open-Next" button to just move onto the next image. This may not be as elegant as filters (makes the operator the filter which also may introduce bias) but if there is a reasonably finite number of images might be the fastest and maybe even more reliable. > > --- Mike > > -----Original Message----- > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Karsten Rodenacker > Sent: Tuesday, March 06, 2007 9:03 AM > To: [hidden email] > Subject: Re: Focus measurement (more general image analysis question) > > Tscha, maybe I am thinking to complicated... > > Possibly you need only a smoothing filter and look than for the difference (again SD and mean). A large SD shows that the original and the smoothed one are different, hence in focus, a small one show the opposite. From that it should be possible to deduce the rule for decision. For smoothing you can apply Smooth or FFT/inverse FFT or Graymorphology Open or Close or Median or ... All is dependent on the blurring. If it is simple, symmetrical the linear operators are best suited. If it is not simple the non-linear morphological/rank order ops are good. They are sensitive either for bright sparks (open) or for dark sparks (close). > > Regards > Karsten > > Am 06.03.2007 um 14:41 schrieb Antje: > >> Hi Karsten, >> >> I don't know whether I understood... (maybe I made a diffuse >> description)... I don't want to filter objects within one image, I >> would like to judge whether the majority of objects is in focus or >> not. In general, I get images which are blurred at all or which have a >> nice focus over the whole area. I just would like to analyze the >> images which have well focused objects. And I have to automate this. >> That means, I'd like something like an "out-of-focus" filter over a >> set of several thousand images, which delivers only the images with a >> good focus... >> >> Antje >> >> >> >> Karsten Rodenacker schrieb: >>> Ok, now we come to some progress, hopefully. >>> You have the detected cell/nuclei, than >>> -apply the already recommended Sobel or any other edge detector to >>> the vesicle fluorescence image, >>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>> (don't ask me how that is done, but I know it is possible) >>> -find the threshold(s) for which YOU consider the vesicle as bad >>> focussed >>> -throw away the objects falling under the criteria >>> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >>> Regards >>> Karsten >>> Am 06.03.2007 um 11:42 schrieb Antje: >>>> Hi Karsten, >>>> >>>> okay, I try to explain what I would like to do. >>>> Let's assume I have 2 images (gray-values) of different >>>> fluorochromes of the same objects. On is representing the nucleus >>>> and the cell, the other represents a dotty structure within the >>>> cell (vesicles). >>>> Now, I would like to detect the cell, the nucleus and the >>>> vesicles, so that I can measure them (e.g. size of the vesicles). >>>> But I would like to exclude that I get biased data from objects >>>> which are bad focused. A blurred vesicle then is much larger, >>>> than a sharp one, though they might have the same size in reality... >>>> My aim is to do high content analysis in an automated way (maybe >>>> also extract further parameters of the vesicles and to feed some >>>> kind of classifier with the data). >>>> >>>> Ciao, >>>> Antje >>>> >>>> >>>> >>>> Karsten Rodenacker schrieb: >>>>> Hi Antje, >>>>> a quite difficult discussion track. Perhaps you should explain >>>>> what you would like to do. I tried to say that there is no focus >>>>> per se. You should describe for what which focus is necessary. >>>>> If you like to estimate maker density per cell number or cell >>>>> volume the focus might be not so important. Since number >>>>> estimation is expensive (time consuming) possibly the density >>>>> per volume is sufficient. I have had very often discussions >>>>> about necessity! Unbiased Stereology from Howard and Reed might >>>>> help. >>>>> Regards >>>>> Karsten >>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>> Hi Jan, >>>>>> >>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>> values over the whole image will be influenced by the number of >>>>>> objects (density), I guess. Further, object detection before, >>>>>> might be dependent on the focus so that you could end up with >>>>>> values you cannot compare. >>>>>> Is there any other way to use the gradient information? >>>>>> >>>>>> Ciao, >>>>>> Antje >>>>>> >>>>>> >>>>>> >>>>>> Jan Eglinger schrieb: >>>>>>> Dear Antje, >>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>>> "Find Edges") and to take the brightness of the result as a >>>>>>> score. >>>>>>> Images with sharp edges should produce a high score that >>>>>>> permits you to sort your images. >>>>>>> I don't know if this will work with your set of images, but it >>>>>>> should be a quick way to get some idea of the "focus" quality >>>>>>> (provided your objects are mainly in a single plane of focus). >>>>>>> Best, >>>>>>> Jan >>>>>>> -------- Original Message -------- >>>>>>> From: Antje <[hidden email]> >>>>>>> To: [hidden email] >>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>> question) >>>>>>> Date: 06.03.2007 10:04 >>>>>>>> Hi Karsten, >>>>>>>> >>>>>>>> you're right confocal microscopic images should be in focus >>>>>>>> by definition. It's for sure that our microscope needs to >>>>>>>> undergo improvements... >>>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>>> have to cope with out of focus images (unfortunately, I'm not >>>>>>>> that deep into physics to be able to explain the reason for >>>>>>>> this focus problems). >>>>>>>> I don't have stacks. >>>>>>>> >>>>>>>> Ciao, >>>>>>>> Antje >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>>> always meant for an object, adjust focus is meant for a >>>>>>>>> whole image. Hence the problem with focus is the observer, a >>>>>>>>> cell of certain thickness might be in focus at the cell >>>>>>>>> border or at the cell center. >>>>>>>>> >>>>>>>>> However, still there is a more important question. Images >>>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>>> images in 'bad focus' maybe there is some help in microscope >>>>>>>>> adjustment necessary? >>>>>>>>> If you have stacks of images, images taken at different >>>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>>> However, my first remark is still valid, meaning the outcome >>>>>>>>> of such a method is possibly far away from expectation. >>>>>>>>> Regards >>>>>>>>> Karsten >>>>>>>>> >>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>> >>>>>>>>>> Hello, >>>>>>>>>> >>>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>>> >>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>> automatically filter images which have a bad focus. First, >>>>>>>>>> I was thinking about some gradient values to judge but they >>>>>>>>>> are so dependent of the content (number of cells... , >>>>>>>>>> presence of the marker). Is there any reliable method known >>>>>>>>>> to measure and compare the focus of images??? Or does >>>>>>>>>> anybody know papers dealing with this issue??? >>>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>>> but I could not find anything...) >>>>>>>>>> >>>>>>>>>> Antje >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ___________________________________________________________ >>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>>> messenger.yahoo.de >>>>>>>> >>>>>>>> >>>>>>>> ___________________________________________________________ >>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>> messenger.yahoo.de >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> ___________________________________________________________ Der >>>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>>> Mail: http://mail.yahoo.de >>>> >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>> messenger.yahoo.de >> >> >> ___________________________________________________________ >> Telefonate ohne weitere Kosten vom PC zum PC: http:// >> messenger.yahoo.de > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
Re: Focus measurement (more general image analysis question)
|
|
In reply to this post by Antje-2
On Wednesday 07 March 2007 13:10:27 Antje wrote:
> But still I have a question. How shall I compare > images with different density? Because the standard deviation of an > image will be dependent on the density of objects. It may happen, that > there is just one cell within one image and it can also happen, that > there is no background at all because of the density of cells... I think it is not possible to get a robust method of sorting *single* images according to their degree of "in-focusness". The assumption of an absolute measure of sharpness for a single image most probably would not hold unless you know already what to expect in the image. If you look/google/search autofocus, you will find that all focusing algorithms try to maximise some measure of sharpness across several shots of the *same scene*. With a single arbitrary image (as I believe is your case), how do you make sure that the image is blurry because of bad focus rather than the scene having originally no sharp edges. I.e. is this 1) a blurry shot of a sharp scene or 2) is it a sharp shot of a diffuse-looking scene? If you look for high frequency contents in the image to make the decision, you would treat the 2 examples above as the same, while an autofocus algorithm would find the best solutions for both cases (by taking more shots at various focal lengths). You then could compare which one was more out of focus (how far away each image was from the "best focus" shot). But you cannot do this with a single shot of unknown properties when "in focus". Cheers, G. |
|
|
In reply to this post by Antje-2
... in principle you are right. However normally the large area
without cells looks also different in or out of focus! Gabriel's posting just arrived is a continuation. However my assumption is that it should be possible to see focal properties also in the background! Gabriel is completely right. There is no solution for your problem only some suboptimal possibilities. It refers back to my first posting: focus is a very subjective property! Regards Karsten Am 07.03.2007 um 14:45 schrieb Antje: > Hi Karsten, > > I'm afraid, I still did not get it. > Lets assume, I have an image with only one single cell (in focus). > Then, of course the difference image shows up quite "high" values > for this area. But calculating a mean or a stddev will be biased by > the large area having low values (because there is nothing inside). > Am I wrong??? > It will return different values than an image containing a lot of > cells with the same focus), no? > > which other posts do you mean? > > ciao, > Antje > > > > Karsten Rodenacker schrieb: >> ... the difference between the image and the smoothed version, >> that is all what is smoothed away, should not be too dependent on >> intensity! The idea is to apply an operation which is similar to >> defocussing. The more the original looks like a smoothed one the >> more the image is out of focus AND the difference will have a >> smaller deviation. Of course a smooth original without edges and >> hence without objects cannot be considered as in or out of focus. >> Have a look also into the other posts! >> My first idea tended to classify cellular objects, my second one >> for whole images, if only some objects are out of focus than it >> has to be decided if this image is acceptable or not. Global >> (image based) measures will smooth local focal properties! >> Regards >> Karsten >> Am 07.03.2007 um 14:10 schrieb Antje: >>> Hi Karsten, >>> >>> that's a nice idea. But still I have a question. How shall I >>> compare images with different density? Because the standard >>> deviation of an image will be dependent on the density of >>> objects. It may happen, that there is just one cell within one >>> image and it can also happen, that there is no background at all >>> because of the density of cells... >>> Or did I miss something of your idea? >>> Every value I can think of seems to be dependent of the content/ >>> background ratio... >>> >>> Ciao, >>> Antje >>> >>> >>> Karsten Rodenacker schrieb: >>>> Tscha, maybe I am thinking to complicated... >>>> Possibly you need only a smoothing filter and look than for the >>>> difference (again SD and mean). A large SD shows that the >>>> original and the smoothed one are different, hence in focus, a >>>> small one show the opposite. From that it should be possible to >>>> deduce the rule for decision. For smoothing you can apply Smooth >>>> or FFT/inverse FFT or Graymorphology Open or Close or Median >>>> or ... All is dependent on the blurring. If it is simple, >>>> symmetrical the linear operators are best suited. If it is not >>>> simple the non-linear morphological/rank order ops are good. >>>> They are sensitive either for bright sparks (open) or for dark >>>> sparks (close). >>>> Regards >>>> Karsten >>>> Am 06.03.2007 um 14:41 schrieb Antje: >>>>> Hi Karsten, >>>>> >>>>> I don't know whether I understood... (maybe I made a diffuse >>>>> description)... I don't want to filter objects within one >>>>> image, I would like to judge whether the majority of objects is >>>>> in focus or not. In general, I get images which are blurred at >>>>> all or which have a nice focus over the whole area. I just >>>>> would like to analyze the images which have well focused >>>>> objects. And I have to automate this. That means, I'd like >>>>> something like an "out-of-focus" filter over a set of several >>>>> thousand images, which delivers only the images with a good >>>>> focus... >>>>> >>>>> Antje >>>>> >>>>> >>>>> >>>>> Karsten Rodenacker schrieb: >>>>>> Ok, now we come to some progress, hopefully. >>>>>> You have the detected cell/nuclei, than >>>>>> -apply the already recommended Sobel or any other edge >>>>>> detector to the vesicle fluorescence image, >>>>>> -Measure SD and mean Gray Value (Redirect To the filtered >>>>>> image) (don't ask me how that is done, but I know it is possible) >>>>>> -find the threshold(s) for which YOU consider the vesicle as >>>>>> bad focussed >>>>>> -throw away the objects falling under the criteria >>>>>> typically SD or SD/Mean can discriminate unfocussed (smoothed) >>>>>> areas. >>>>>> Regards >>>>>> Karsten >>>>>> Am 06.03.2007 um 11:42 schrieb Antje: >>>>>>> Hi Karsten, >>>>>>> >>>>>>> okay, I try to explain what I would like to do. >>>>>>> Let's assume I have 2 images (gray-values) of different >>>>>>> fluorochromes of the same objects. On is representing the >>>>>>> nucleus and the cell, the other represents a dotty structure >>>>>>> within the cell (vesicles). >>>>>>> Now, I would like to detect the cell, the nucleus and the >>>>>>> vesicles, so that I can measure them (e.g. size of the >>>>>>> vesicles). >>>>>>> But I would like to exclude that I get biased data from >>>>>>> objects which are bad focused. A blurred vesicle then is much >>>>>>> larger, than a sharp one, though they might have the same >>>>>>> size in reality... >>>>>>> My aim is to do high content analysis in an automated way >>>>>>> (maybe also extract further parameters of the vesicles and to >>>>>>> feed some kind of classifier with the data). >>>>>>> >>>>>>> Ciao, >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> >>>>>>> Karsten Rodenacker schrieb: >>>>>>>> Hi Antje, >>>>>>>> a quite difficult discussion track. Perhaps you should >>>>>>>> explain what you would like to do. I tried to say that there >>>>>>>> is no focus per se. You should describe for what which focus >>>>>>>> is necessary. >>>>>>>> If you like to estimate maker density per cell number or >>>>>>>> cell volume the focus might be not so important. Since >>>>>>>> number estimation is expensive (time consuming) possibly the >>>>>>>> density per volume is sufficient. I have had very often >>>>>>>> discussions about necessity! Unbiased Stereology from Howard >>>>>>>> and Reed might help. >>>>>>>> Regards >>>>>>>> Karsten >>>>>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>>>>> Hi Jan, >>>>>>>>> >>>>>>>>> if I take a Sobel filter, it'll be dependent of the amount >>>>>>>>> of objects within the image. Mean, Median or sum of the >>>>>>>>> gradient values over the whole image will be influenced by >>>>>>>>> the number of objects (density), I guess. Further, object >>>>>>>>> detection before, might be dependent on the focus so that >>>>>>>>> you could end up with values you cannot compare. >>>>>>>>> Is there any other way to use the gradient information? >>>>>>>>> >>>>>>>>> Ciao, >>>>>>>>> Antje >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Jan Eglinger schrieb: >>>>>>>>>> Dear Antje, >>>>>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel >>>>>>>>>> filter "Find Edges") and to take the brightness of the >>>>>>>>>> result as a score. >>>>>>>>>> Images with sharp edges should produce a high score that >>>>>>>>>> permits you to sort your images. >>>>>>>>>> I don't know if this will work with your set of images, >>>>>>>>>> but it should be a quick way to get some idea of the >>>>>>>>>> "focus" quality (provided your objects are mainly in a >>>>>>>>>> single plane of focus). >>>>>>>>>> Best, >>>>>>>>>> Jan >>>>>>>>>> -------- Original Message -------- >>>>>>>>>> From: Antje <[hidden email]> >>>>>>>>>> To: [hidden email] >>>>>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>>>>> question) >>>>>>>>>> Date: 06.03.2007 10:04 >>>>>>>>>>> Hi Karsten, >>>>>>>>>>> >>>>>>>>>>> you're right confocal microscopic images should be in >>>>>>>>>>> focus by definition. It's for sure that our microscope >>>>>>>>>>> needs to undergo improvements... >>>>>>>>>>> Anyhow, due to a large screening mode of this microscope, >>>>>>>>>>> I have to cope with out of focus images (unfortunately, >>>>>>>>>>> I'm not that deep into physics to be able to explain the >>>>>>>>>>> reason for this focus problems). >>>>>>>>>>> I don't have stacks. >>>>>>>>>>> >>>>>>>>>>> Ciao, >>>>>>>>>>> Antje >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>>>>> You are tackling an interesting question. To be in focus >>>>>>>>>>>> is always meant for an object, adjust focus is meant for >>>>>>>>>>>> a whole image. Hence the problem with focus is the >>>>>>>>>>>> observer, a cell of certain thickness might be in focus >>>>>>>>>>>> at the cell border or at the cell center. >>>>>>>>>>>> >>>>>>>>>>>> However, still there is a more important question. >>>>>>>>>>>> Images from confocal microscopes are by definition in >>>>>>>>>>>> focus of course related to the 'pin-hole' size. If you >>>>>>>>>>>> call such images in 'bad focus' maybe there is some help >>>>>>>>>>>> in microscope adjustment necessary? >>>>>>>>>>>> If you have stacks of images, images taken at different >>>>>>>>>>>> focal adjustments there is a program to merge such >>>>>>>>>>>> stacks into one image (e.g. plugin Exteded depth of >>>>>>>>>>>> field). However, my first remark is still valid, meaning >>>>>>>>>>>> the outcome of such a method is possibly far away from >>>>>>>>>>>> expectation. >>>>>>>>>>>> Regards >>>>>>>>>>>> Karsten >>>>>>>>>>>> >>>>>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>>>>> >>>>>>>>>>>>> Hello, >>>>>>>>>>>>> >>>>>>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>>>>>> >>>>>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>>>>> automatically filter images which have a bad focus. >>>>>>>>>>>>> First, I was thinking about some gradient values to >>>>>>>>>>>>> judge but they are so dependent of the content (number >>>>>>>>>>>>> of cells... , presence of the marker). Is there any >>>>>>>>>>>>> reliable method known to measure and compare the focus >>>>>>>>>>>>> of images??? Or does anybody know papers dealing with >>>>>>>>>>>>> this issue??? >>>>>>>>>>>>> (If there is something available in ImageJ it would be >>>>>>>>>>>>> fine but I could not find anything...) >>>>>>>>>>>>> >>>>>>>>>>>>> Antje >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> __________________________________________________________ >>>>>>>>>>>>> _ Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>>>>>> messenger.yahoo.de >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>>>>>> messenger.yahoo.de >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ___________________________________________________________ >>>>>>>>> Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen >>>>>>>>> Yahoo! Mail: http://mail.yahoo.de >>>>>>> >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>>>> messenger.yahoo.de >>>>> >>>>> >>>>> >>>>> ___________________________________________________________ >>>>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>>>> messenger.yahoo.de >>> >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http:// >>> messenger.yahoo.de > > > > > > > ___________________________________________________________ Der > frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! > Mail: http://mail.yahoo.de |
|
|
In reply to this post by Gabriel Landini
Hi Gabriel,
In general you are completely right! But I don't have the possibility to get more than one shot per scene. That's a fact! I was hoping that there is some kind of measurement which can judge on "sharpness" for images with different content independently from the content. So, that I can get values which are comparable between images with different densities of cells e.g. Then, I can have a look at these values and decide where to put the threshold to judge if something is in focus or not. But therefore, I need some kind of content independent measurement... I can assume that all images have the same kind of content (sharp scenes of nuclei for example) and now I'd like to compare these images not taking into account the amount of "content". Maybe it is not possible (I don't wanna believe... ) But sooner or later I have to overcome this problem... Antje Gabriel Landini schrieb: > On Wednesday 07 March 2007 13:10:27 Antje wrote: >> But still I have a question. How shall I compare >> images with different density? Because the standard deviation of an >> image will be dependent on the density of objects. It may happen, that >> there is just one cell within one image and it can also happen, that >> there is no background at all because of the density of cells... > > I think it is not possible to get a robust method of sorting > *single* images according to their degree of "in-focusness". > > The assumption of an absolute measure of sharpness for a single image most > probably would not hold unless you know already what to expect in the image. > > If you look/google/search autofocus, you will find that all focusing > algorithms try to maximise some measure of sharpness across several shots of > the *same scene*. > With a single arbitrary image (as I believe is your case), how do you make > sure that the image is blurry because of bad focus rather than the scene > having originally no sharp edges. > I.e. is this 1) a blurry shot of a sharp scene or 2) is it a sharp shot of a > diffuse-looking scene? > > If you look for high frequency contents in the image to make the decision, you > would treat the 2 examples above as the same, while an autofocus algorithm > would find the best solutions for both cases (by taking more shots at various > focal lengths). You then could compare which one was more out of focus (how > far away each image was from the "best focus" shot). > But you cannot do this with a single shot of unknown properties when "in > focus". > > Cheers, > G. > ___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de |
|
|
In reply to this post by karo03
Okay :) I see,
at least I'm sure not to miss a common/simple method to solve this problem by asking you for your opinion. I'll investigate the background thing you mentioned. Maybe you are right and the difference is also there. Thank you for your patience! Antje Karsten Rodenacker schrieb: > ... in principle you are right. However normally the large area without > cells looks also different in or out of focus! > Gabriel's posting just arrived is a continuation. > However my assumption is that it should be possible to see focal > properties also in the background! > > Gabriel is completely right. There is no solution for your problem only > some suboptimal possibilities. It refers back to my first posting: focus > is a very subjective property! > > Regards > Karsten > > Am 07.03.2007 um 14:45 schrieb Antje: > >> Hi Karsten, >> >> I'm afraid, I still did not get it. >> Lets assume, I have an image with only one single cell (in focus). >> Then, of course the difference image shows up quite "high" values for >> this area. But calculating a mean or a stddev will be biased by the >> large area having low values (because there is nothing inside). Am I >> wrong??? >> It will return different values than an image containing a lot of >> cells with the same focus), no? >> >> which other posts do you mean? >> >> ciao, >> Antje >> >> >> >> Karsten Rodenacker schrieb: >>> ... the difference between the image and the smoothed version, that >>> is all what is smoothed away, should not be too dependent on >>> intensity! The idea is to apply an operation which is similar to >>> defocussing. The more the original looks like a smoothed one the more >>> the image is out of focus AND the difference will have a smaller >>> deviation. Of course a smooth original without edges and hence >>> without objects cannot be considered as in or out of focus. >>> Have a look also into the other posts! >>> My first idea tended to classify cellular objects, my second one for >>> whole images, if only some objects are out of focus than it has to be >>> decided if this image is acceptable or not. Global (image based) >>> measures will smooth local focal properties! >>> Regards >>> Karsten >>> Am 07.03.2007 um 14:10 schrieb Antje: >>>> Hi Karsten, >>>> >>>> that's a nice idea. But still I have a question. How shall I compare >>>> images with different density? Because the standard deviation of an >>>> image will be dependent on the density of objects. It may happen, >>>> that there is just one cell within one image and it can also happen, >>>> that there is no background at all because of the density of cells... >>>> Or did I miss something of your idea? >>>> Every value I can think of seems to be dependent of the >>>> content/background ratio... >>>> >>>> Ciao, >>>> Antje >>>> >>>> >>>> Karsten Rodenacker schrieb: >>>>> Tscha, maybe I am thinking to complicated... >>>>> Possibly you need only a smoothing filter and look than for the >>>>> difference (again SD and mean). A large SD shows that the original >>>>> and the smoothed one are different, hence in focus, a small one >>>>> show the opposite. From that it should be possible to deduce the >>>>> rule for decision. For smoothing you can apply Smooth or >>>>> FFT/inverse FFT or Graymorphology Open or Close or Median or ... >>>>> All is dependent on the blurring. If it is simple, symmetrical the >>>>> linear operators are best suited. If it is not simple the >>>>> non-linear morphological/rank order ops are good. They are >>>>> sensitive either for bright sparks (open) or for dark sparks (close). >>>>> Regards >>>>> Karsten >>>>> Am 06.03.2007 um 14:41 schrieb Antje: >>>>>> Hi Karsten, >>>>>> >>>>>> I don't know whether I understood... (maybe I made a diffuse >>>>>> description)... I don't want to filter objects within one image, I >>>>>> would like to judge whether the majority of objects is in focus or >>>>>> not. In general, I get images which are blurred at all or which >>>>>> have a nice focus over the whole area. I just would like to >>>>>> analyze the images which have well focused objects. And I have to >>>>>> automate this. That means, I'd like something like an >>>>>> "out-of-focus" filter over a set of several thousand images, which >>>>>> delivers only the images with a good focus... >>>>>> >>>>>> Antje >>>>>> >>>>>> >>>>>> >>>>>> Karsten Rodenacker schrieb: >>>>>>> Ok, now we come to some progress, hopefully. >>>>>>> You have the detected cell/nuclei, than >>>>>>> -apply the already recommended Sobel or any other edge detector >>>>>>> to the vesicle fluorescence image, >>>>>>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>>>>>> (don't ask me how that is done, but I know it is possible) >>>>>>> -find the threshold(s) for which YOU consider the vesicle as bad >>>>>>> focussed >>>>>>> -throw away the objects falling under the criteria >>>>>>> typically SD or SD/Mean can discriminate unfocussed (smoothed) >>>>>>> areas. >>>>>>> Regards >>>>>>> Karsten >>>>>>> Am 06.03.2007 um 11:42 schrieb Antje: >>>>>>>> Hi Karsten, >>>>>>>> >>>>>>>> okay, I try to explain what I would like to do. >>>>>>>> Let's assume I have 2 images (gray-values) of different >>>>>>>> fluorochromes of the same objects. On is representing the >>>>>>>> nucleus and the cell, the other represents a dotty structure >>>>>>>> within the cell (vesicles). >>>>>>>> Now, I would like to detect the cell, the nucleus and the >>>>>>>> vesicles, so that I can measure them (e.g. size of the vesicles). >>>>>>>> But I would like to exclude that I get biased data from objects >>>>>>>> which are bad focused. A blurred vesicle then is much larger, >>>>>>>> than a sharp one, though they might have the same size in >>>>>>>> reality... >>>>>>>> My aim is to do high content analysis in an automated way (maybe >>>>>>>> also extract further parameters of the vesicles and to feed some >>>>>>>> kind of classifier with the data). >>>>>>>> >>>>>>>> Ciao, >>>>>>>> Antje >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>> Hi Antje, >>>>>>>>> a quite difficult discussion track. Perhaps you should explain >>>>>>>>> what you would like to do. I tried to say that there is no >>>>>>>>> focus per se. You should describe for what which focus is >>>>>>>>> necessary. >>>>>>>>> If you like to estimate maker density per cell number or cell >>>>>>>>> volume the focus might be not so important. Since number >>>>>>>>> estimation is expensive (time consuming) possibly the density >>>>>>>>> per volume is sufficient. I have had very often discussions >>>>>>>>> about necessity! Unbiased Stereology from Howard and Reed might >>>>>>>>> help. >>>>>>>>> Regards >>>>>>>>> Karsten >>>>>>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>>>>>> Hi Jan, >>>>>>>>>> >>>>>>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>>>>>> values over the whole image will be influenced by the number >>>>>>>>>> of objects (density), I guess. Further, object detection >>>>>>>>>> before, might be dependent on the focus so that you could end >>>>>>>>>> up with values you cannot compare. >>>>>>>>>> Is there any other way to use the gradient information? >>>>>>>>>> >>>>>>>>>> Ciao, >>>>>>>>>> Antje >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Jan Eglinger schrieb: >>>>>>>>>>> Dear Antje, >>>>>>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel >>>>>>>>>>> filter "Find Edges") and to take the brightness of the result >>>>>>>>>>> as a score. >>>>>>>>>>> Images with sharp edges should produce a high score that >>>>>>>>>>> permits you to sort your images. >>>>>>>>>>> I don't know if this will work with your set of images, but >>>>>>>>>>> it should be a quick way to get some idea of the "focus" >>>>>>>>>>> quality (provided your objects are mainly in a single plane >>>>>>>>>>> of focus). >>>>>>>>>>> Best, >>>>>>>>>>> Jan >>>>>>>>>>> -------- Original Message -------- >>>>>>>>>>> From: Antje <[hidden email]> >>>>>>>>>>> To: [hidden email] >>>>>>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>>>>>> question) >>>>>>>>>>> Date: 06.03.2007 10:04 >>>>>>>>>>>> Hi Karsten, >>>>>>>>>>>> >>>>>>>>>>>> you're right confocal microscopic images should be in focus >>>>>>>>>>>> by definition. It's for sure that our microscope needs to >>>>>>>>>>>> undergo improvements... >>>>>>>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>>>>>>> have to cope with out of focus images (unfortunately, I'm >>>>>>>>>>>> not that deep into physics to be able to explain the reason >>>>>>>>>>>> for this focus problems). >>>>>>>>>>>> I don't have stacks. >>>>>>>>>>>> >>>>>>>>>>>> Ciao, >>>>>>>>>>>> Antje >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>>>>>>> always meant for an object, adjust focus is meant for a >>>>>>>>>>>>> whole image. Hence the problem with focus is the observer, >>>>>>>>>>>>> a cell of certain thickness might be in focus at the cell >>>>>>>>>>>>> border or at the cell center. >>>>>>>>>>>>> >>>>>>>>>>>>> However, still there is a more important question. Images >>>>>>>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>>>>>>> images in 'bad focus' maybe there is some help in >>>>>>>>>>>>> microscope adjustment necessary? >>>>>>>>>>>>> If you have stacks of images, images taken at different >>>>>>>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>>>>>>> However, my first remark is still valid, meaning the >>>>>>>>>>>>> outcome of such a method is possibly far away from >>>>>>>>>>>>> expectation. >>>>>>>>>>>>> Regards >>>>>>>>>>>>> Karsten >>>>>>>>>>>>> >>>>>>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>>>>>> >>>>>>>>>>>>>> Hello, >>>>>>>>>>>>>> >>>>>>>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>>>>>>> >>>>>>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>>>>>> automatically filter images which have a bad focus. First, >>>>>>>>>>>>>> I was thinking about some gradient values to judge but >>>>>>>>>>>>>> they are so dependent of the content (number of cells... , >>>>>>>>>>>>>> presence of the marker). Is there any reliable method >>>>>>>>>>>>>> known to measure and compare the focus of images??? Or >>>>>>>>>>>>>> does anybody know papers dealing with this issue??? >>>>>>>>>>>>>> (If there is something available in ImageJ it would be >>>>>>>>>>>>>> fine but I could not find anything...) >>>>>>>>>>>>>> >>>>>>>>>>>>>> Antje >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>>>>>> http://messenger.yahoo.de >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>>>> http://messenger.yahoo.de >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> ___________________________________________________________ >>>>>>>>>> Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen >>>>>>>>>> Yahoo! Mail: http://mail.yahoo.de >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ___________________________________________________________ >>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>> http://messenger.yahoo.de >>>>>> >>>>>> >>>>>> ___________________________________________________________ >>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>> http://messenger.yahoo.de >>>> >>>> >>>> ___________________________________________________________ >>>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> >> >> >> >> >> >> ___________________________________________________________ Der frühe >> Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: >> http://mail.yahoo.de > ___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de |
Re: Focus measurement (more general image analysis question)
|
|
In reply to this post by Antje-2
If you have the image with the nucleus labeling, could you just count
the number of cells present in an image (very basic analyze particles) and use it to do some kind of normalization, or at least compensate the effect of density to calculate the "focusness index" of the whole set (nuclei + vesicles + whatever channel you have) ? Christophe Antje a écrit : > Hi Karsten, > > I'm afraid, I still did not get it. > Lets assume, I have an image with only one single cell (in focus). Then, > of course the difference image shows up quite "high" values for this > area. But calculating a mean or a stddev will be biased by the large > area having low values (because there is nothing inside). Am I wrong??? > It will return different values than an image containing a lot of cells > with the same focus), no? > > which other posts do you mean? > > ciao, > Antje > > > > Karsten Rodenacker schrieb: >> ... the difference between the image and the smoothed version, that is >> all what is smoothed away, should not be too dependent on intensity! >> The idea is to apply an operation which is similar to defocussing. The >> more the original looks like a smoothed one the more the image is out >> of focus AND the difference will have a smaller deviation. Of course a >> smooth original without edges and hence without objects cannot be >> considered as in or out of focus. >> >> Have a look also into the other posts! >> >> My first idea tended to classify cellular objects, my second one for >> whole images, if only some objects are out of focus than it has to be >> decided if this image is acceptable or not. Global (image based) >> measures will smooth local focal properties! >> Regards >> Karsten >> >> Am 07.03.2007 um 14:10 schrieb Antje: >> >>> Hi Karsten, >>> >>> that's a nice idea. But still I have a question. How shall I compare >>> images with different density? Because the standard deviation of an >>> image will be dependent on the density of objects. It may happen, >>> that there is just one cell within one image and it can also happen, >>> that there is no background at all because of the density of cells... >>> Or did I miss something of your idea? >>> Every value I can think of seems to be dependent of the >>> content/background ratio... >>> >>> Ciao, >>> Antje >>> >>> >>> Karsten Rodenacker schrieb: >>>> Tscha, maybe I am thinking to complicated... >>>> Possibly you need only a smoothing filter and look than for the >>>> difference (again SD and mean). A large SD shows that the original >>>> and the smoothed one are different, hence in focus, a small one show >>>> the opposite. From that it should be possible to deduce the rule for >>>> decision. For smoothing you can apply Smooth or FFT/inverse FFT or >>>> Graymorphology Open or Close or Median or ... All is dependent on >>>> the blurring. If it is simple, symmetrical the linear operators are >>>> best suited. If it is not simple the non-linear morphological/rank >>>> order ops are good. They are sensitive either for bright sparks >>>> (open) or for dark sparks (close). >>>> Regards >>>> Karsten >>>> Am 06.03.2007 um 14:41 schrieb Antje: >>>>> Hi Karsten, >>>>> >>>>> I don't know whether I understood... (maybe I made a diffuse >>>>> description)... I don't want to filter objects within one image, I >>>>> would like to judge whether the majority of objects is in focus or >>>>> not. In general, I get images which are blurred at all or which >>>>> have a nice focus over the whole area. I just would like to analyze >>>>> the images which have well focused objects. And I have to automate >>>>> this. That means, I'd like something like an "out-of-focus" filter >>>>> over a set of several thousand images, which delivers only the >>>>> images with a good focus... >>>>> >>>>> Antje >>>>> >>>>> >>>>> >>>>> Karsten Rodenacker schrieb: >>>>>> Ok, now we come to some progress, hopefully. >>>>>> You have the detected cell/nuclei, than >>>>>> -apply the already recommended Sobel or any other edge detector to >>>>>> the vesicle fluorescence image, >>>>>> -Measure SD and mean Gray Value (Redirect To the filtered image) >>>>>> (don't ask me how that is done, but I know it is possible) >>>>>> -find the threshold(s) for which YOU consider the vesicle as bad >>>>>> focussed >>>>>> -throw away the objects falling under the criteria >>>>>> typically SD or SD/Mean can discriminate unfocussed (smoothed) areas. >>>>>> Regards >>>>>> Karsten >>>>>> Am 06.03.2007 um 11:42 schrieb Antje: >>>>>>> Hi Karsten, >>>>>>> >>>>>>> okay, I try to explain what I would like to do. >>>>>>> Let's assume I have 2 images (gray-values) of different >>>>>>> fluorochromes of the same objects. On is representing the nucleus >>>>>>> and the cell, the other represents a dotty structure within the >>>>>>> cell (vesicles). >>>>>>> Now, I would like to detect the cell, the nucleus and the >>>>>>> vesicles, so that I can measure them (e.g. size of the vesicles). >>>>>>> But I would like to exclude that I get biased data from objects >>>>>>> which are bad focused. A blurred vesicle then is much larger, >>>>>>> than a sharp one, though they might have the same size in reality... >>>>>>> My aim is to do high content analysis in an automated way (maybe >>>>>>> also extract further parameters of the vesicles and to feed some >>>>>>> kind of classifier with the data). >>>>>>> >>>>>>> Ciao, >>>>>>> Antje >>>>>>> >>>>>>> >>>>>>> >>>>>>> Karsten Rodenacker schrieb: >>>>>>>> Hi Antje, >>>>>>>> a quite difficult discussion track. Perhaps you should explain >>>>>>>> what you would like to do. I tried to say that there is no focus >>>>>>>> per se. You should describe for what which focus is necessary. >>>>>>>> If you like to estimate maker density per cell number or cell >>>>>>>> volume the focus might be not so important. Since number >>>>>>>> estimation is expensive (time consuming) possibly the density >>>>>>>> per volume is sufficient. I have had very often discussions >>>>>>>> about necessity! Unbiased Stereology from Howard and Reed might >>>>>>>> help. >>>>>>>> Regards >>>>>>>> Karsten >>>>>>>> Am 06.03.2007 um 10:51 schrieb Antje: >>>>>>>>> Hi Jan, >>>>>>>>> >>>>>>>>> if I take a Sobel filter, it'll be dependent of the amount of >>>>>>>>> objects within the image. Mean, Median or sum of the gradient >>>>>>>>> values over the whole image will be influenced by the number of >>>>>>>>> objects (density), I guess. Further, object detection before, >>>>>>>>> might be dependent on the focus so that you could end up with >>>>>>>>> values you cannot compare. >>>>>>>>> Is there any other way to use the gradient information? >>>>>>>>> >>>>>>>>> Ciao, >>>>>>>>> Antje >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Jan Eglinger schrieb: >>>>>>>>>> Dear Antje, >>>>>>>>>> I'd suggest to use an edge detection filter (e.g. Sobel filter >>>>>>>>>> "Find Edges") and to take the brightness of the result as a >>>>>>>>>> score. >>>>>>>>>> Images with sharp edges should produce a high score that >>>>>>>>>> permits you to sort your images. >>>>>>>>>> I don't know if this will work with your set of images, but it >>>>>>>>>> should be a quick way to get some idea of the "focus" quality >>>>>>>>>> (provided your objects are mainly in a single plane of focus). >>>>>>>>>> Best, >>>>>>>>>> Jan >>>>>>>>>> -------- Original Message -------- >>>>>>>>>> From: Antje <[hidden email]> >>>>>>>>>> To: [hidden email] >>>>>>>>>> Subject: Re:Focus measurement (more general image analysis >>>>>>>>>> question) >>>>>>>>>> Date: 06.03.2007 10:04 >>>>>>>>>>> Hi Karsten, >>>>>>>>>>> >>>>>>>>>>> you're right confocal microscopic images should be in focus >>>>>>>>>>> by definition. It's for sure that our microscope needs to >>>>>>>>>>> undergo improvements... >>>>>>>>>>> Anyhow, due to a large screening mode of this microscope, I >>>>>>>>>>> have to cope with out of focus images (unfortunately, I'm not >>>>>>>>>>> that deep into physics to be able to explain the reason for >>>>>>>>>>> this focus problems). >>>>>>>>>>> I don't have stacks. >>>>>>>>>>> >>>>>>>>>>> Ciao, >>>>>>>>>>> Antje >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Karsten Rodenacker schrieb: >>>>>>>>>>>> You are tackling an interesting question. To be in focus is >>>>>>>>>>>> always meant for an object, adjust focus is meant for a >>>>>>>>>>>> whole image. Hence the problem with focus is the observer, a >>>>>>>>>>>> cell of certain thickness might be in focus at the cell >>>>>>>>>>>> border or at the cell center. >>>>>>>>>>>> >>>>>>>>>>>> However, still there is a more important question. Images >>>>>>>>>>>> from confocal microscopes are by definition in focus of >>>>>>>>>>>> course related to the 'pin-hole' size. If you call such >>>>>>>>>>>> images in 'bad focus' maybe there is some help in microscope >>>>>>>>>>>> adjustment necessary? >>>>>>>>>>>> If you have stacks of images, images taken at different >>>>>>>>>>>> focal adjustments there is a program to merge such stacks >>>>>>>>>>>> into one image (e.g. plugin Exteded depth of field). >>>>>>>>>>>> However, my first remark is still valid, meaning the outcome >>>>>>>>>>>> of such a method is possibly far away from expectation. >>>>>>>>>>>> Regards >>>>>>>>>>>> Karsten >>>>>>>>>>>> >>>>>>>>>>>> Am 06.03.2007 um 09:17 schrieb Antje: >>>>>>>>>>>> >>>>>>>>>>>>> Hello, >>>>>>>>>>>>> >>>>>>>>>>>>> sorry, the question is not that strongly related with >>>>>>>>>>>>> ImageJ but maybe there is still someone who can help? >>>>>>>>>>>>> >>>>>>>>>>>>> I have a huge set of microscopic images from a confocal >>>>>>>>>>>>> microscope (cells/nuclei + marker) and I'd like to >>>>>>>>>>>>> automatically filter images which have a bad focus. First, >>>>>>>>>>>>> I was thinking about some gradient values to judge but they >>>>>>>>>>>>> are so dependent of the content (number of cells... , >>>>>>>>>>>>> presence of the marker). Is there any reliable method known >>>>>>>>>>>>> to measure and compare the focus of images??? Or does >>>>>>>>>>>>> anybody know papers dealing with this issue??? >>>>>>>>>>>>> (If there is something available in ImageJ it would be fine >>>>>>>>>>>>> but I could not find anything...) >>>>>>>>>>>>> >>>>>>>>>>>>> Antje >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>>>>> http://messenger.yahoo.de >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> ___________________________________________________________ >>>>>>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>>>>>> http://messenger.yahoo.de >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ___________________________________________________________ Der >>>>>>>>> frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! >>>>>>>>> Mail: http://mail.yahoo.de >>>>>>> >>>>>>> >>>>>>> >>>>>>> ___________________________________________________________ >>>>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>>>> http://messenger.yahoo.de >>>>> >>>>> >>>>> ___________________________________________________________ >>>>> Telefonate ohne weitere Kosten vom PC zum PC: >>>>> http://messenger.yahoo.de >>> >>> >>> ___________________________________________________________ >>> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de >> > > > > > > > ___________________________________________________________ Der frühe > Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: > http://mail.yahoo.de > |
«
Return to ImageJ
|
1 view|%1 views
| Free forum by Nabble | Edit this page |