ImageJ

Segmentation Issue to resolve

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

4 messages Options

anusuya pal

Segmentation Issue to resolve

Hello,

I am very new in this field and I am wondering if somebody can help me to
figure out about the segmentation of each domain. I have converted the
image into 8bit and I want to separate the different region of interest to
find the mean intensity for each region. For convenience, I am attaching
the image.

Thanks
Anu

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

0.jpg (145K) Download Attachment

ctrueden

Re: Segmentation Issue to resolve

Hi Anu,

I tried a few strategies with your image, but it seems very difficult. I
was not able to produce a good segmentation with Trainable Weka
Segmentation [1] nor with Morphological Segmentation [2]. The latter seems
more promising than the former, but the contrast between your small
interior regions and larger regions along the perimeter makes a unified
segmentation very challenging.

[image: Inline image 2][image: Inline image 1]

You could try posting on the ImageJ Forum (http://forum.imagej.net/); maybe
someone there has further suggestions or workflows for you to try.

Regards,
Curtis

[1] https://imagej.net/TWS
[2] https://imagej.net/Morphological_Segmentation

--
Curtis Rueden
LOCI software architect - https://loci.wisc.edu/software
ImageJ2 lead, Fiji maintainer - https://imagej.net/User:Rueden
Did you know ImageJ has a forum? http://forum.imagej.net/

On Fri, Sep 8, 2017 at 6:08 PM, anusuya pal <[hidden email]> wrote:

> Hello,
>
> I am very new in this field and I am wondering if somebody can help me to
> figure out about the segmentation of each domain. I have converted the
> image into 8bit and I want to separate the different region of interest to
> find the mean intensity for each region. For convenience, I am attaching
> the image.
>
> Thanks
> Anu
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

morphological-segmentation.png (1M) Download Attachment

tws.png (1M) Download Attachment

anusuya pal

Re: Segmentation Issue to resolve

Thanks a lot. I have tried both the segmentation but it didn't work well.
So, I am thinking in a different direction and have given a new post. I
just hope that will be easier to solve.

Thanks
Anu

On 11-Sep-2017 2:19 PM, "Curtis Rueden" <[hidden email]> wrote:

Hi Anu,

I tried a few strategies with your image, but it seems very difficult. I
was not able to produce a good segmentation with Trainable Weka
Segmentation [1] nor with Morphological Segmentation [2]. The latter seems
more promising than the former, but the contrast between your small
interior regions and larger regions along the perimeter makes a unified
segmentation very challenging.

[image: Inline image 2][image: Inline image 1]

You could try posting on the ImageJ Forum (http://forum.imagej.net/); maybe
someone there has further suggestions or workflows for you to try.

Regards,
Curtis

[1] https://imagej.net/TWS
[2] https://imagej.net/Morphological_Segmentation

--
Curtis Rueden
LOCI software architect - https://loci.wisc.edu/software
ImageJ2 lead, Fiji maintainer - https://imagej.net/User:Rueden
Did you know ImageJ has a forum? http://forum.imagej.net/

On Fri, Sep 8, 2017 at 6:08 PM, anusuya pal <[hidden email]> wrote:

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

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

anusuya pal

Re: Segmentation Issue to resolve

Hello,

I have an image to do the segmentation. As the image is very complex, the
basic segmentations (Trainable Weka Segmentation, Morphological
Segmentation, water shedding with BIovoxxel tool etc) can't be used
accurately. Is there any method so that I can only do the segmentation for
the big and the small domains (i.e. ignoring the holes, just to get the
border of each domain)? For the convenience, I am attaching the raw image
(converted into 8 bit) and the image containing the region of interest.

Here is the link
https://drive.google.com/drive/folders/0B-zIvUHKCb_zOUgtSnI2d1pnRHM?usp=sharing

Thanks
Anu

On Mon, Sep 11, 2017 at 2:38 PM, anusuya pal <[hidden email]> wrote:

> Thanks a lot. I have tried both the segmentation but it didn't work well.
> So, I am thinking in a different direction and have given a new post. I
> just hope that will be easier to solve.
>
> Thanks
> Anu
>
>
>
> On 11-Sep-2017 2:19 PM, "Curtis Rueden" <[hidden email]> wrote:
>
> Hi Anu,
>
> I tried a few strategies with your image, but it seems very difficult. I
> was not able to produce a good segmentation with Trainable Weka
> Segmentation [1] nor with Morphological Segmentation [2]. The latter seems
> more promising than the former, but the contrast between your small
> interior regions and larger regions along the perimeter makes a unified
> segmentation very challenging.
>
> [image: Inline image 2][image: Inline image 1]
>
> You could try posting on the ImageJ Forum (http://forum.imagej.net/);
> maybe
> someone there has further suggestions or workflows for you to try.
>
> Regards,
> Curtis
>
> [1] https://imagej.net/TWS
> [2] https://imagej.net/Morphological_Segmentation
>
> --
> Curtis Rueden
> LOCI software architect - https://loci.wisc.edu/software
> ImageJ2 lead, Fiji maintainer - https://imagej.net/User:Rueden
> Did you know ImageJ has a forum? http://forum.imagej.net/
>
>
> On Fri, Sep 8, 2017 at 6:08 PM, anusuya pal <[hidden email]>
> wrote:
>
> > Hello,
> >
> > I am very new in this field and I am wondering if somebody can help me to
> > figure out about the segmentation of each domain. I have converted the
> > image into 8bit and I want to separate the different region of interest
> to
> > find the mean intensity for each region. For convenience, I am attaching
> > the image.
> >
> > Thanks
> > Anu
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
>
>

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