Hi all,

I am working on particle analysis and I am trying to automate cell count. I came up with an
algorithm which first filters the noise , then does segmentation, closes  unclosed cells, fills
holes and then does  cell count. The algorithm works well for some images, however there are
problems when this code is applied to images in which the cells are of the same clarity as the
background. Segmentation removes cells or part of the cells that are of the same contrast as
the background. I want to improve my algorithm to make sure all the cells are correctly
identified   even if they are merged with the background or better still if the cells can be
adaptively segmented  from  the background. In other words, I want a robust algorithm that
works well with any image with good or bad clarity.

Any suggestions will help me proceed further.

Thanks,
Sandhya