I have an image A with object O1, and another image B with two objects O1 but scaled and O2. I looke for a similarity measure taking into acount the changes of scale. I tried to use histogram intersection because I thaugth that is robust to scale, but it seems that it doesn't work. And now I am thinking about Fourier transform but I am not sure about its properties, and if it is useful in my case (scale changes + occlusions). I put into your hand an example of 2 images.
Try SSIM (structural similarity).
https://en.wikipedia.org/wiki/Structural_similarity
Tanks for your answer, But actually I look for another similarity measure robust to scale . I Have already read about SSIM, Image substruct and I used also correlation combined with the variance between this two images but this is useful if there is no scale changes either it doesn't work.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images