Pattern recognition algorithm to for detecting blood cells [closed]

Question

Closed. This question needs to be more focused. It is not currently accepting answers.

---

Want to improve this question? Update the question so it focuses on one problem only by editing this post.

Closed 8 years ago.

Improve this question

I have a following type of image (blood cells):

Is there any pattern recognition algorithm, which can detect blood cells?

Answer 1

An OpenCV library includes a Hough algorithm, which allows to find circles in image.

Answer 2

In the image you posted, it looks like the strongest signals are: color of the red blood cells, their edges, and the fact that they're blobs of roughly the same size.

The simple algorithm below might work. You can google how to do each of these things in OpenCV or MATLAB or whatever you're using.

1. Make a binary image where white = (the pixel is close to red blood cell color), black = (the pixel is far from red blood cell color). Experiment with the threshold until it looks good.

2. Blur the image, and then count local maxima (with the radius of blurring and of "local maxima" being the expected blood cell radius, or a bit smaller). The blurring serves to put a white peak in the center of a blob.

If the above is not good enough, you may need to do a more advanced method based on the fact that strong brightness gradients tend to point away from the red blood cell center. The pdf in my pupil-detection Github project discusses how to do that: https://github.com/LukeAllen/optimeyes

Answer 3

Check out this paper...

Automatic red blood cell counting using hough transform

Link