Object recognition in an image?

Question

Image processing

Hello. I have a large number of pictures of cells with different shapes and I need to count the number of cells of each shape in the picture. Processing all this manually is long and tedious, plus it became interesting whether there are programs that ask the image of the object to be found, and it looks for it in other images. Tell me what programs (free) are there to solve this problem.
I am attaching an example for clarity:

Answer 1

There is most likely no ready-made solution for cell classification - because the task is quite specific.
However, with proper experience, it is quite solvable with the help of existing computer vision algorithms that are publicly available.
I recommend looking towards OpenCV.
Actually, your specific task falls into two sub-tasks:
1) Detection of cells.
2) Their classification.
On the first question, apparently, you can either use a circle detector (all cells have a shape that is approximately the same round in the first approximation), or train some detector from OpenCV (HAAR, LBP or HOG) to detect them.
On the second point - to analyze the properties of cells, choose the characteristics for which their difference is the greatest and train some classifier (For example, SVM).

Answer 2

I can recommend this filter for ImageMagick:
convert $1 -define convolve:scale=! -morphology Convolve Disk:5 -morphology EdgeIn Octagon -negate -normalize -threshold 80% -morphology Erode Disk:2 $2
For example, this picture is obtained.
Most of the cells are detected by a white area of ​​a certain size, similar to a circle. The similarity to a circle can be calculated by the ratio of the distance from the center of mass to the farthest and closest points on the border.
True, sometimes broken cells are obtained ... (see upper left and right corners)