I'm trying to find the best algorithm for my task. I have images looking like this one:

enter image description here

I want to find the most similar ones from a set of such images. I tried several algorithms to compute the similarity value (sift - opencv2, structural similarities - skimage.measure.compare_ssim, pixel similarity, wasserstein_distance...). None of them seems to work as I'd like to. For example sift returned almost maximal similarity for the above image and this one: enter image description here

Which is far from perfect. Are there any algorithms which could perform better on such a task?

  • $\begingroup$ You say you want the most similar but you don't define what you mean by similarity or how you propose to measure it. Therefore, I don't think the question is answerable. As you have discovered, there are many possible ways to measure similarity, none of them perfect. Without giving us your preferred measure of similarity or some properties which it must satisfy, I don't think the question is answerable. "Better" is a matter of opinion and "such a task" is vague. $\endgroup$ – D.W. Apr 13 at 16:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.