3D Suite - 3D Distance map, file input bug?

[Matthew Jones]

Hi,

Apologies if my last email was mangled - the text seemed to disappear.

I get different results depending on which image I process with the 3D distance map plugin. The metadata for the two files is different, however I don’t know when this difference originated (possibly when using an older version of imageJ?).

Is anyone able to tell what is different about the images from the metadata and how to convert from ‘bad’ to ‘good’ format.

Also, a big thank you to Thomas Boudier for 3D Suite and related plugins as they have been invaluable in my research and generally work excellently!

Kind regards,

Matt

p.s. my system is  Macbook Pro OS X Yosemite 10.10.3 and FIJI (though some of the images may have been processed in an older version of ImageJ at an earlier stage).
original files at: https://drive.google.com/open?id=0B_9C1WFBZYoUbzRBVE5nUXJrWDg

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Thomas Boudier-4]

Hi Matt,

The problem is the calibration of the image, for the "good" one it´s
0.06x0.06x0.4 microns and for the "bad" one it´s 0.16x0.16x0.16, so it
looks like the Z calibration is lost for the "bad" one. You can easily
check if the calibration is correct by visualising your data in 3D viewer .

Best,

Thomas

On 19/07/2017 23:57, Matthew Jones wrote:

> Hi,
>
> Apologies if my last email was mangled - the text seemed to disappear.
>
> I get different results depending on which image I process with the 3D distance map plugin. The metadata for the two files is different, however I don’t know when this difference originated (possibly when using an older version of imageJ?).
>
> Is anyone able to tell what is different about the images from the metadata and how to convert from ‘bad’ to ‘good’ format.
>
> Also, a big thank you to Thomas Boudier for 3D Suite and related plugins as they have been invaluable in my research and generally work excellently!
>
> Kind regards,
>
> Matt
>
> p.s. my system is  Macbook Pro OS X Yosemite 10.10.3 and FIJI (though some of the images may have been processed in an older version of ImageJ at an earlier stage).
> original files at: https://drive.google.com/open?id=0B_9C1WFBZYoUbzRBVE5nUXJrWDg
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
  /***************************************************************/
       Thomas Boudier, Associate Professor, UPMC,
       Université Pierre et Marie Curie, Paris, France.
       BioInformatics Institute (BII)/IPAL, Singapore.
/**************************************************************/


--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Michael Doube-4]

Hi Matt,

Thomas raises an important point about 3D distance maps which is that in some implementations (e.g. in Bob Dougherty's Local Thickness plugin), the pixels are treated as though they have isotropic spacing, regardless of true pixel spacing.

That could make a mess of your results if your z spacing is different to your x and y spacing, which is the usual case for confocal microscopy and clinical CT images.

You can get around it by interpolating/rescaling in z so that z spacing matches x and y spacing, but this risks inventing data where there wasn't any collected.

Michael


On 20/07/17 03:17, Thomas Boudier wrote:

> Hi Matt,
>
> The problem is the calibration of the image, for the "good" one it´s 0.06x0.06x0.4 microns and for the "bad" one it´s 0.16x0.16x0.16, so it looks like the Z calibration is lost for the "bad" one. You can easily check if the calibration is correct by visualising your data in 3D viewer .
>
> Best,
>
> Thomas
>
> On 19/07/2017 23:57, Matthew Jones wrote:
>> Hi,
>>
>> Apologies if my last email was mangled - the text seemed to disappear.
>>
>> I get different results depending on which image I process with the 3D distance map plugin. The metadata for the two files is different, however I don’t know when this difference originated (possibly when using an older version of imageJ?).
>>
>> Is anyone able to tell what is different about the images from the metadata and how to convert from ‘bad’ to ‘good’ format.
>>
>> Also, a big thank you to Thomas Boudier for 3D Suite and related plugins as they have been invaluable in my research and generally work excellently!
>>
>> Kind regards,
>>
>> Matt
>>
>> p.s. my system is  Macbook Pro OS X Yosemite 10.10.3 and FIJI (though some of the images may have been processed in an older version of ImageJ at an earlier stage).
>> original files at: https://drive.google.com/open?id=0B_9C1WFBZYoUbzRBVE5nUXJrWDg
>>
>> --
>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
Michael Doube, PhD
Lecturer, Comparative Biomedical Sciences
The Royal Veterinary College, University of London
Royal College Street
London NW1 0TU
United Kingdom

[phone] (Internal: 5503)
@mdoube

<http://www.rvc.ac.uk>

This message, together with any attachments, is intended for the stated addressee(s) only and may contain privileged or confidential information. Any views or opinions presented are solely those of the author and do not necessarily represent those of the Royal Veterinary College.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Matthew Jones]

Thank you Thomas and Michael,

I’m really grateful for your help thanks for pointing out this potential pitfall in my analysis regarding the difference between x/y and z voxel dimensions.

I have z spacing of 2 micron in one cell type (0.16micron x and y), and 0.4 micron in the other cell type (0.06 x and y).

Would either of you advise any interpolation or rescaling?  The attached image is one of the 2 micron spaced segmented images. The middle of the object doesn’t look too bad, but the top and bottom look a bit unusual.

I use the distance maps of this space to look up the location of an object detected in another channel, so none of my points fall outside the map, but I appreciate that my map might be distorted by low sampling in the z direction.

The attached volume viewer screenshot suggests the distance map looks OK in cross section.

Any thoughts or suggestions for improving this analysis would be welcome,

thanks,

Matt

> On 20 Jul 2017, at 09:43, Michael Doube <[hidden email]> wrote:
>
> Hi Matt,
>
> Thomas raises an important point about 3D distance maps which is that in some implementations (e.g. in Bob Dougherty's Local Thickness plugin), the pixels are treated as though they have isotropic spacing, regardless of true pixel spacing.
>
> That could make a mess of your results if your z spacing is different to your x and y spacing, which is the usual case for confocal microscopy and clinical CT images.
>
> You can get around it by interpolating/rescaling in z so that z spacing matches x and y spacing, but this risks inventing data where there wasn't any collected.
>
> Michael
>
>
> On 20/07/17 03:17, Thomas Boudier wrote:
>> Hi Matt,
>>
>> The problem is the calibration of the image, for the "good" one it´s 0.06x0.06x0.4 microns and for the "bad" one it´s 0.16x0.16x0.16, so it looks like the Z calibration is lost for the "bad" one. You can easily check if the calibration is correct by visualising your data in 3D viewer .
>>
>> Best,
>>
>> Thomas
>>
>> On 19/07/2017 23:57, Matthew Jones wrote:
>>> Hi,
>>>
>>> Apologies if my last email was mangled - the text seemed to disappear.
>>>
>>> I get different results depending on which image I process with the 3D distance map plugin. The metadata for the two files is different, however I don’t know when this difference originated (possibly when using an older version of imageJ?).
>>>
>>> Is anyone able to tell what is different about the images from the metadata and how to convert from ‘bad’ to ‘good’ format.
>>>
>>> Also, a big thank you to Thomas Boudier for 3D Suite and related plugins as they have been invaluable in my research and generally work excellently!
>>>
>>> Kind regards,
>>>
>>> Matt
>>>
>>> p.s. my system is  Macbook Pro OS X Yosemite 10.10.3 and FIJI (though some of the images may have been processed in an older version of ImageJ at an earlier stage).
>>> original files at: https://drive.google.com/open?id=0B_9C1WFBZYoUbzRBVE5nUXJrWDg
>>>
>>> --
>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>
>
>
> --
> Michael Doube, PhD
> Lecturer, Comparative Biomedical Sciences
> The Royal Veterinary College, University of London
> Royal College Street
> London NW1 0TU
> United Kingdom
>
> [phone] (Internal: 5503)
> @mdoube
>
> <http://www.rvc.ac.uk <http://www.rvc.ac.uk/>>
>
> This message, together with any attachments, is intended for the stated addressee(s) only and may contain privileged or confidential information. Any views or opinions presented are solely those of the author and do not necessarily represent those of the Royal Veterinary College.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html <http://imagej.nih.gov/ij/list.html>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html