mem leak II
Locked
 
|
I, too, am getting memory leakage and running out of 1135mb mem on my pc(XP) using last month's Ij version.
I'm analysing more than 2k images, cycling each 110mb rsb tiff: finding center of region, cropping, thresholding, dilating 10 times each 16mb bw tif and running analyze particles to counts. Simple,straight forward stuff. I've monitored and see the 2-3 mb mem leak after each iteration cycle. I've inserted into my macro six 'call("java....gc");' after each set of 50, but to no avail. Not using setBatchMode. Any advice? Thanks, Bill R. |
|
|
Hi,
I think there is a bug in the cell counter plugin. I had the problem in imageJ v1.44f and it also occurs in v1.44g. To recreate the problem... 1. open any image (organ of corti will do) and initialise the cell counter plugin on the window 2. add a cell type beyond type 8... ie just add one more so you have 9 types 3. click a few types with types 1-8... click a few times with type 9 4. save markers 5. close > reopen organ of corti > re-initialise the cell counter plugin on the window > load markers When this is done... only types 1-8 are present in the cell counter plugin... they all display 0 despite the fact that you clicked in the original... moreover type 9 is not loaded when you load the markers. If you click "add"... hey presto... all the numbers are then there. Can others recreate this? Mike -------------------------------------------------- From: "Rothman, William" <[hidden email]> Sent: Friday, September 17, 2010 10:07 PM To: <[hidden email]> Subject: mem leak II > I, too, am getting memory leakage and running out of 1135mb mem on my > pc(XP) using last month's Ij version. > I'm analysing more than 2k images, cycling each 110mb rsb tiff: finding > center of region, cropping, thresholding, dilating 10 times each 16mb bw > tif and running analyze particles to counts. > Simple,straight forward stuff. > I've monitored and see the 2-3 mb mem leak after each iteration cycle. > I've inserted into my macro six 'call("java....gc");' after each set of > 50, but to no avail. Not using setBatchMode. > Any advice? > Thanks, > Bill R. > > > __________ Information from ESET Smart Security, version of virus > signature database 5458 (20100917) __________ > > The message was checked by ESET Smart Security. > > http://www.eset.com > > > __________ Information from ESET Smart Security, version of virus signature database 5463 (20100920) __________ The message was checked by ESET Smart Security. http://www.eset.com |
|
|
Im using imageJ v1.44g
Can others recreate this bug: 1. open organ of corti from samples 2. image > colour > split channels 3. image > colour > channels tool > more > make composite > set display mode to greyscale > hit ok The new composite image doesnt display in grey... its still colour. If you have the channels tool still open then click greyscale... the image goes grey fine as it should of done before. However, if you close the channels tool, then reopen the channels tool... theres like 7 channels listed. This bug happens with all my own images. Im using LOCI bioformats importer to import deltavision files, I check the split channels function within LOCI bioformats and set display mode to greyscale but when the channels split they are always in colour. Is this mailing list the right way to report potential bugs? Thanks Mike -------------------------------------------------- From: "Michael Cundell" <[hidden email]> Sent: Monday, September 20, 2010 12:23 PM To: <[hidden email]> Subject: cell counter bug... > Hi, > > I think there is a bug in the cell counter plugin. I had the problem in > imageJ v1.44f and it also occurs in v1.44g. > > To recreate the problem... > 1. open any image (organ of corti will do) and initialise the cell counter > plugin on the window > 2. add a cell type beyond type 8... ie just add one more so you have 9 > types > 3. click a few types with types 1-8... click a few times with type 9 > 4. save markers > 5. close > reopen organ of corti > re-initialise the cell counter plugin > on the window > load markers > > When this is done... only types 1-8 are present in the cell counter > plugin... they all display 0 despite the fact that you clicked in the > original... moreover type 9 is not loaded when you load the markers. If > you click "add"... hey presto... all the numbers are then there. > > Can others recreate this? > > Mike > > > -------------------------------------------------- > From: "Rothman, William" <[hidden email]> > Sent: Friday, September 17, 2010 10:07 PM > To: <[hidden email]> > Subject: mem leak II > >> I, too, am getting memory leakage and running out of 1135mb mem on my >> pc(XP) using last month's Ij version. >> I'm analysing more than 2k images, cycling each 110mb rsb tiff: finding >> center of region, cropping, thresholding, dilating 10 times each 16mb bw >> tif and running analyze particles to counts. >> Simple,straight forward stuff. >> I've monitored and see the 2-3 mb mem leak after each iteration cycle. >> I've inserted into my macro six 'call("java....gc");' after each set of >> 50, but to no avail. Not using setBatchMode. >> Any advice? >> Thanks, >> Bill R. >> >> >> __________ Information from ESET Smart Security, version of virus >> signature database 5458 (20100917) __________ >> >> The message was checked by ESET Smart Security. >> >> http://www.eset.com >> >> >> > > __________ Information from ESET Smart Security, version of virus > signature database 5463 (20100920) __________ > > The message was checked by ESET Smart Security. > > http://www.eset.com > > > __________ Information from ESET Smart Security, version of virus > signature database 5464 (20100920) __________ > > The message was checked by ESET Smart Security. > > http://www.eset.com > > > __________ Information from ESET Smart Security, version of virus signature database 5468 (20100921) __________ The message was checked by ESET Smart Security. http://www.eset.com |
Re: composite > grayscale bug.
|
|
Hi Mike (and everyone else),
> This bug happens with all my own images. Im using LOCI bioformats > importer to import deltavision files, I check the split channels > function within LOCI bioformats and set display mode to greyscale > but when the channels split they are always in colour. This problem has been fixed in the very latest trunk build of Bio-Formats. You should be able to upgrade within ImageJ by selecting "Plugins > LOCI > Update LOCI Plugins", or manually by following the directions here: http://loci.wisc.edu/bio-formats/imagej. For anyone who is tracking the 4.2 branch of Bio-Formats, this was also fixed in SVN revision 6972 on the 4.2 branch. Regards, -Melissa On Wed, Sep 22, 2010 at 09:12:54AM +0100, Michael Cundell wrote: > Im using imageJ v1.44g > > Can others recreate this bug: > 1. open organ of corti from samples > 2. image > colour > split channels > 3. image > colour > channels tool > more > make composite > set > display mode to greyscale > hit ok > > The new composite image doesnt display in grey... its still colour. > If you have the channels tool still open then click greyscale... the > image goes grey fine as it should of done before. However, if you > close the channels tool, then reopen the channels tool... theres > like 7 channels listed. > > This bug happens with all my own images. Im using LOCI bioformats > importer to import deltavision files, I check the split channels > function within LOCI bioformats and set display mode to greyscale > but when the channels split they are always in colour. > > Is this mailing list the right way to report potential bugs? > > Thanks > > Mike > > -------------------------------------------------- > From: "Michael Cundell" <[hidden email]> > Sent: Monday, September 20, 2010 12:23 PM > To: <[hidden email]> > Subject: cell counter bug... > > >Hi, > > > >I think there is a bug in the cell counter plugin. I had the > >problem in imageJ v1.44f and it also occurs in v1.44g. > > > >To recreate the problem... > >1. open any image (organ of corti will do) and initialise the cell > >counter plugin on the window > >2. add a cell type beyond type 8... ie just add one more so you > >have 9 types > >3. click a few types with types 1-8... click a few times with type 9 > >4. save markers > >5. close > reopen organ of corti > re-initialise the cell counter > >plugin on the window > load markers > > > >When this is done... only types 1-8 are present in the cell > >counter plugin... they all display 0 despite the fact that you > >clicked in the original... moreover type 9 is not loaded when you > >load the markers. If you click "add"... hey presto... all the > >numbers are then there. > > > >Can others recreate this? > > > >Mike > > > > > >-------------------------------------------------- > >From: "Rothman, William" <[hidden email]> > >Sent: Friday, September 17, 2010 10:07 PM > >To: <[hidden email]> > >Subject: mem leak II > > > >>I, too, am getting memory leakage and running out of 1135mb mem > >>on my pc(XP) using last month's Ij version. > >>I'm analysing more than 2k images, cycling each 110mb rsb tiff: > >>finding center of region, cropping, thresholding, dilating 10 > >>times each 16mb bw tif and running analyze particles to counts. > >>Simple,straight forward stuff. > >>I've monitored and see the 2-3 mb mem leak after each iteration cycle. > >>I've inserted into my macro six 'call("java....gc");' after > >>each set of 50, but to no avail. Not using setBatchMode. > >>Any advice? > >>Thanks, > >>Bill R. > >> > >> > >>__________ Information from ESET Smart Security, version of > >>virus signature database 5458 (20100917) __________ > >> > >>The message was checked by ESET Smart Security. > >> > >>http://www.eset.com > >> > >> > >> > > > >__________ Information from ESET Smart Security, version of virus > >signature database 5463 (20100920) __________ > > > >The message was checked by ESET Smart Security. > > > >http://www.eset.com > > > > > >__________ Information from ESET Smart Security, version of virus > >signature database 5464 (20100920) __________ > > > >The message was checked by ESET Smart Security. > > > >http://www.eset.com > > > > > > > > __________ Information from ESET Smart Security, version of virus signature database 5468 (20100921) __________ > > The message was checked by ESET Smart Security. > > http://www.eset.com |
Re: composite > grayscale bug.
|
|
On Wednesday 22 Sep 2010 09:23:12 Melissa Linkert wrote:
> You should be able to upgrade within ImageJ by selecting > "Plugins > LOCI > Update LOCI Plugins", or manually by following the > directions here: http://loci.wisc.edu/bio-formats/imagej. Hi Melissa, Just a related question to the above. The "Update LOCI Plugins" works fine, but at the end, it asks to restart IJ. Is that really necessary? There is a 'Help>Refresh Menus' and I wondered whether the installer could call directly this command. Regards Gabriel |
|
|
Hi,
On Thu, 23 Sep 2010, Gabriel Landini wrote: > On Wednesday 22 Sep 2010 09:23:12 Melissa Linkert wrote: > > You should be able to upgrade within ImageJ by selecting "Plugins > > > LOCI > Update LOCI Plugins", or manually by following the directions > > here: http://loci.wisc.edu/bio-formats/imagej. > > Just a related question to the above. The "Update LOCI Plugins" works > fine, but at the end, it asks to restart IJ. Is that really necessary? > There is a 'Help>Refresh Menus' and I wondered whether the installer > could call directly this command. The problem here is not so much that the new classes would not be read. The problem is that old classes from that .jar might still be in use. Just think about opening a virtual stack via Bio-Formats and then updating the LOCI plugins. Sure, most of the time the buffers will contain the bytes from the old .jar and all goes fine. But that is by chance, not by design. Once those buffers were flushed, when Java tries to access the instance it reads the new bytes from the new .jar (which are likely not the same as the old ones), and all kind of hell breaks loose. Just as an example: when I overwrite my ij.jar (and make sure that the relevant buffer is flushed), the typical symptom is that I cannot even quit anymore, because it does not find the "Quit" action. I have to kill ImageJ. The "Refresh Menus" is a cute hack, but a hack nevertheless. Ciao, Dscho |
|
|
In reply to this post by Melissa Linkert-2
Thanks Melissa.
Mike -------------------------------------------------- From: "Melissa Linkert" <[hidden email]> Sent: Wednesday, September 22, 2010 5:15 PM To: <[hidden email]> Subject: Re: composite > grayscale bug. > Hi Mike (and everyone else), > >> This bug happens with all my own images. Im using LOCI bioformats >> importer to import deltavision files, I check the split channels >> function within LOCI bioformats and set display mode to greyscale >> but when the channels split they are always in colour. > > This problem has been fixed in the very latest trunk build of > Bio-Formats. You should be able to upgrade within ImageJ by selecting > "Plugins > LOCI > Update LOCI Plugins", or manually by following the > directions here: http://loci.wisc.edu/bio-formats/imagej. > > For anyone who is tracking the 4.2 branch of Bio-Formats, this was also > fixed in SVN revision 6972 on the 4.2 branch. > > Regards, > -Melissa > > On Wed, Sep 22, 2010 at 09:12:54AM +0100, Michael Cundell wrote: >> Im using imageJ v1.44g >> >> Can others recreate this bug: >> 1. open organ of corti from samples >> 2. image > colour > split channels >> 3. image > colour > channels tool > more > make composite > set >> display mode to greyscale > hit ok >> >> The new composite image doesnt display in grey... its still colour. >> If you have the channels tool still open then click greyscale... the >> image goes grey fine as it should of done before. However, if you >> close the channels tool, then reopen the channels tool... theres >> like 7 channels listed. >> >> This bug happens with all my own images. Im using LOCI bioformats >> importer to import deltavision files, I check the split channels >> function within LOCI bioformats and set display mode to greyscale >> but when the channels split they are always in colour. >> >> Is this mailing list the right way to report potential bugs? >> >> Thanks >> >> Mike >> >> -------------------------------------------------- >> From: "Michael Cundell" <[hidden email]> >> Sent: Monday, September 20, 2010 12:23 PM >> To: <[hidden email]> >> Subject: cell counter bug... >> >> >Hi, >> > >> >I think there is a bug in the cell counter plugin. I had the >> >problem in imageJ v1.44f and it also occurs in v1.44g. >> > >> >To recreate the problem... >> >1. open any image (organ of corti will do) and initialise the cell >> >counter plugin on the window >> >2. add a cell type beyond type 8... ie just add one more so you >> >have 9 types >> >3. click a few types with types 1-8... click a few times with type 9 >> >4. save markers >> >5. close > reopen organ of corti > re-initialise the cell counter >> >plugin on the window > load markers >> > >> >When this is done... only types 1-8 are present in the cell >> >counter plugin... they all display 0 despite the fact that you >> >clicked in the original... moreover type 9 is not loaded when you >> >load the markers. If you click "add"... hey presto... all the >> >numbers are then there. >> > >> >Can others recreate this? >> > >> >Mike >> > >> > >> >-------------------------------------------------- >> >From: "Rothman, William" <[hidden email]> >> >Sent: Friday, September 17, 2010 10:07 PM >> >To: <[hidden email]> >> >Subject: mem leak II >> > >> >>I, too, am getting memory leakage and running out of 1135mb mem >> >>on my pc(XP) using last month's Ij version. >> >>I'm analysing more than 2k images, cycling each 110mb rsb tiff: >> >>finding center of region, cropping, thresholding, dilating 10 >> >>times each 16mb bw tif and running analyze particles to counts. >> >>Simple,straight forward stuff. >> >>I've monitored and see the 2-3 mb mem leak after each iteration cycle. >> >>I've inserted into my macro six 'call("java....gc");' after >> >>each set of 50, but to no avail. Not using setBatchMode. >> >>Any advice? >> >>Thanks, >> >>Bill R. >> >> >> >> >> >>__________ Information from ESET Smart Security, version of >> >>virus signature database 5458 (20100917) __________ >> >> >> >>The message was checked by ESET Smart Security. >> >> >> >>http://www.eset.com >> >> >> >> >> >> >> > >> >__________ Information from ESET Smart Security, version of virus >> >signature database 5463 (20100920) __________ >> > >> >The message was checked by ESET Smart Security. >> > >> >http://www.eset.com >> > >> > >> >__________ Information from ESET Smart Security, version of virus >> >signature database 5464 (20100920) __________ >> > >> >The message was checked by ESET Smart Security. >> > >> >http://www.eset.com >> > >> > >> > >> >> __________ Information from ESET Smart Security, version of virus >> signature database 5468 (20100921) __________ >> >> The message was checked by ESET Smart Security. >> >> http://www.eset.com > > > __________ Information from ESET Smart Security, version of virus > signature database 5470 (20100922) __________ > > The message was checked by ESET Smart Security. > > http://www.eset.com > > > __________ Information from ESET Smart Security, version of virus signature database 5471 (20100922) __________ The message was checked by ESET Smart Security. http://www.eset.com |
«
Return to ImageJ
|
1 view|%1 views
| Free forum by Nabble | Edit this page |