Mosaic creation of 2 images using Harris Corners, SIFT descriptors and RANSAC
This code uses VLFeat library, for creating the SIFT descriptors from the found Harris features. Instructions for the installation can be found here.
Calculate the SIFT descriptors using the Harris features. Then use vl_ubcmatch() to find the matches. Lastly apply RANSAC to find the similarity transform that fits best. On the left the total matches found by vl_ubcmatch() is shown, and on the right the biggest consensus set that fits the transform found by RANSAC.
Lastly, invert the found transform and fuse two images
Calculate the SIFT descriptors using the Harris features. Then use vl_ubcmatch() to find the matches. Lastly apply RANSAC to find the similarity transform that fits best. On the left the total matches found by vl_ubcmatch() is shown, and on the right the biggest consensus set that fits the transform found by RANSAC.