This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Feature matching means that you want to find similar descriptors. For SIFT, a descriptor vector is a 64 (or 128) dimensional vector and you want to find other vectors with a small distance (in most cases euclidean distance or something similar). FLANN is a library that provides algorithms to find the most similar descriptors out of a large set of descriptors for a given descriptor. To this end it uses a tree-like datastructure that has to be generated before you can find similar descriptors. So in short, you invest some time to create this structure but afterwards, you can find similar descriptors very fast.
