[Insight-users] Fwd: Itk Segmentation Filter for nuclei

[Humayun Irshad]

Thank you Bradley, These images are blue channel of H&E stained images and
you can see color images in mail attachment. I tried
confidenceconnectedImageFilter on blue and
VectorConfidenceConnectedImageFilter on RGB image. But most of time it
crash when i give different values of InitialNeighborhoodRadius and
Multiplier the nuclei. For Blue channel, it work
InitialNeighborhoodRadius = 1 and Multiplier = 2 and for color (RGB) it
work on   InitialNeighborhoodRadius = 2 and Multiplier = 3. There results
are not satisfactory, some time hole inside and some time oversegment. I
also tried geodesic dilation, it doesn't effect the output.

I am not getting suggestion regarding perceptual based color space. Do you
mean HSV color space?

Can you suggest me some filters based Active Contour and level set based
segmentation in ITK?

On Thu, May 3, 2012 at 9:07 PM, Bradley Lowekamp <blowekamp at mail.nih.gov>wrote:

> Hello,
>
> The images you provided are black and white shouldn't H&E stained image be
> color? You may tried using VectorConfidenceConnetedImageFilter, on the
> color image either with a single seed or in groups to estimate the global
> color range of the nuclei. You should also consider the color space used,
> and if it would be best to do the range in a perceptual based space, or not.
>
> If the segmentation is predictable under-segmented by a certain number
> pixels around the boarder you may consider just refining the segmentation
> with a geodesic dilation of an over-segmentation.
>
> Just a couple of suggestions to help you explore algorithms to try to find
> a solution to your problem.
>
> Brad
>
> On May 3, 2012, at 5:46 AM, Humayun Irshad wrote:
>
> ITK Experts,
>
> I need to segment nuclei (specially mitosis nuclei) in H&E stained
> histopathological images (see in attachment). I have seed points. I tried
> ConfidenceConnectedImageFilter and FastMarchingImageFilter but it give poor
> results. Most of the time, nuclei are under segment. Mitosis nuclei
> are homogeneous and having variations in shape and size. Please suggest any
> better segmentation filter.
>
> --
> --
> Best Regards,
>
> *HUMAYUN IRSHAD*
> *
> PhD student, University of Joseph Fourier, Grenoble, France
> **
> Research Engineer,
> **
> IPAL – Image & Pervasive Access Lab, Singapore
> **
> UMI CNRS (I2R/A*STAR, NUS, UJF, UPMC, IT)
>
> Institute for Infocomm Research (I2R)
> 1 Fusionopolis Way
> #10-19 Connexis South Tower
> Singapore 138632
> Tel:+65-65162864
> Cell: +65-83851671
>
> *
>
> <4_B.bmp><0_B.bmp><1_B.bmp><2_B.bmp><3_B.bmp>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.php
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>
>  ========================================================
>
> Bradley Lowekamp
>
> Medical Science and Computing for
>
> Office of High Performance Computing and Communications
>
> National Library of Medicine
>
> blowekamp at mail.nih.gov
>
>
>
>


-- 
-- 
Best Regards,

*HUMAYUN IRSHAD*
*
PhD student, University of Joseph Fourier, Grenoble, France
**
Research Engineer,
**
IPAL – Image & Pervasive Access Lab, Singapore
**
UMI CNRS (I2R/A*STAR, NUS, UJF, UPMC, IT)

Institute for Infocomm Research (I2R)
1 Fusionopolis Way
#10-19 Connexis South Tower
Singapore 138632
Tel:+65-65162864
Cell: +65-83851671

*




-- 
-- 
Best Regards,

*HUMAYUN IRSHAD*
*
PhD student, University of Joseph Fourier, Grenoble, France
**
Research Engineer,
**
IPAL – Image & Pervasive Access Lab, Singapore
**
UMI CNRS (I2R/A*STAR, NUS, UJF, UPMC, IT)

Institute for Infocomm Research (I2R)
1 Fusionopolis Way
#10-19 Connexis South Tower
Singapore 138632
Tel:+65-65162864
Cell: +65-83851671

*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20120504/051572d0/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 6.bmp
Type: image/bmp
Size: 17918 bytes
Desc: not available
URL: <http://www.itk.org/pipermail/insight-users/attachments/20120504/051572d0/attachment-0002.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0.bmp
Type: image/bmp
Size: 17918 bytes
Desc: not available
URL: <http://www.itk.org/pipermail/insight-users/attachments/20120504/051572d0/attachment-0003.bin>