CellProfiler: Identifying objects from enhanced edge picture

Dear All,
I have done some blood smear imaging and I want to process my data using cell profiler for identifying my cells and counting them.
I have ended up with some bad quality images that I could enhance with enhance edge pipeline.
Here is the result of my process and as you can see it could draw the borders of my cells quite well but now unfortunately I am having difficulties in recognizing them ! I was trying to adjust my cell size parameter to be able to detect cells that fall in erythroids range but as you see I have not been successful. Does anyone here have any hints or suggestion to detect my objectd correctly ?


Can you please upload original images. Did you try First Thresholding image then detecting objects using IdentifyPrimaryObjects?


The reason this isn’t working the way you expect is that CellProfiler is set up to find solid light objects on a dark background- which is not what you have here. It’s doing its best, but it therefore comes up with some strange things.

Without knowing more about exactly how your original pipeline works, it’s hard to say exactly what’s the best thing to do from here, but one idea would be to use ImageMath to subtract your Edge image from your Original image, then run Identify on the output from that.