Hi Tom,
I have found in the past that applying a variance filter to enhance
contrast at the edges of the DIC cells works very well. And it also works
on cells with one side dark and the other bright. Then after thresholding
and transforming to binary, simply fill the hole before segmenting.
Sincerely,
*Gabriel Lapointe, M.Sc.*
Lab Manager
Concordia University, Biology Department
7141 Sherbrooke St. West SP 534
Montréal QC H4B 1R6 Canada
Lab : (514) 848-2424 x5988
Fax : (514) 848-2881
Cell : (514) 278-0247
[hidden email]
http://gabriellapointe.ca2011/12/16 Tom Runia <
[hidden email]>
> Hello everybody,
>
> For my bachelor thesis project I am building an automated cell counter in
> ImageJ. For "simple" yeast images at 10 and 60 times magnification this
> works reasonably well.
> Now I am dealing with cells such as these:
http://goo.gl/ja5zh Here the
> cells are much harder to detect because the cells have almost the same
> color as the background.
>
> Does anyone have a good idea or thought about how I can detect and count
> these cells?
> After contrast enhancement, edge finding, and some close/dilate operations
> and a watershed I came up with this:
http://goo.gl/3Im0G> But that probably only works for this special image, and the results is far
> from good...Pretty difficult problem I think...?
>
> Any thoughts are more than welcome :-)
>
> Best,
>
> Tom
>
Locked
