I am working on a program that identifies the most homogeneous circle areas of a specific pixel diameter in the spots of an image (see attached image). Up to now, the program turns the image to gray scale and equalizes the image to have a full 0-255 pixel value range. I believe the next step is to select the black spots (5 of them) as ROI to then iterate through it with a certain step to find the circle of certain diameter that has the least pixel value variation. Is this the way to go? How could I make the program select the 5 spots as ROIs? Thanks
%2011.47.38.png){kind=link}