Segmentation problems

[Girault France]

 
Dear all,
 
I-m trying to segment automatically some particles on images. These particles are not quite different from the background and I was wondering which technique is the most appropriate to apply.I-ve tried different techniques by manipulating the brightness and contrast, blurring the image... but sometimes results are quite good and sometimes it-s a desaster... I cannot find out a generic segmentation technique to apply on my images. Does anyone have some advice or ideas of techniques that I could test on my images ??
Now I-m trying to use the watershed segmentation which works also sometimes only. If I use this plugin, I get several images and one of them called dams gives me the shape that I would like to segment (black and white image) but I cannot grab the form in the middle except whether I convert it to mask, erode, exclude on edge... And by eroding the image, the shape is quite different then.
I was wondering whether there is a better way to grab the shape that appear in this image in order to make a mask...
 
Many thanks
 
france
____________________________
France Girault
Geological Institute
CAB E-62
ETH Zürich
Universitätstrasse 6
8092 Zürich

[hidden email]
Private phone : ++4178/726.99.20
Office phone : ++4144/632 84 29

[Noel BONNET]

Dear France,
Unfortunately (!!!), there is no 'universal' method for image segmentation.
That is why it would be necessary to see one of 'your' images before sending
you an intelligent answer ...
For instance, is your backgroung smooth and of enough constant value? etc ....
If you do not have a lot of particles per image, you could perhaps try the
semi-automatic watershed-based segmentation algorithm we have developed for
ImageJ (http://www.univ-reims.fr/INSERM514/ImageJ).
Although semi-automatic segmentation is much less appealing than fully
automatic image segmentation, it is sometimes rewarding ...
Feel free to send me one of your images, if you want.
Regards.

Noel

[way4thesub]

The perfect solution was just developed at Stanford University - GemIdent. This open-source free package segments cells from miscroscopic images and, more generally, can identify objects of interest and do color image segmentation.

It works particularly well in color images where the objects of interest generally look alike with small variation.

Adam

Girault France wrote

 
Dear all,
 
I-m trying to segment automatically some particles on images. These particles are not quite different from the background and I was wondering which technique is the most appropriate to apply.I-ve tried different techniques by manipulating the brightness and contrast, blurring the image... but sometimes results are quite good and sometimes it-s a desaster... I cannot find out a generic segmentation technique to apply on my images. Does anyone have some advice or ideas of techniques that I could test on my images ??
Now I-m trying to use the watershed segmentation which works also sometimes only. If I use this plugin, I get several images and one of them called dams gives me the shape that I would like to segment (black and white image) but I cannot grab the form in the middle except whether I convert it to mask, erode, exclude on edge... And by eroding the image, the shape is quite different then.
I was wondering whether there is a better way to grab the shape that appear in this image in order to make a mask...
 
Many thanks
 
france
____________________________
France Girault
Geological Institute
CAB E-62
ETH Zürich
Universitätstrasse 6
8092 Zürich

[e-mail]
Private phone : ++4178/726.99.20
Office phone : ++4144/632 84 29