OpenCV Python implementation for finding correspondences between two images


Does anyone know if there is an existing function in OpenCV Python that can calculate the correspondences between two images given a set of the two images and the keypoints detected (no descriptors)? According to papers that calculate the repeatabiltiy score, the correspondences != matches using the descriptors but only the keypoints.

I am aware of the existence of cv2.evaluateFeatureDetectors() but there is not a python wrapper around it yet.

Thank you :slight_smile:

1 Like

You can try:

from skimage.metrics import structural_similarity as compare_ssim
(score, diff) = compare_ssim(img1, img2, full=True)

… if the images are aligned.