VLFeat

VLFeat is an open source software library focused on computer vision, specializing in implementation of state-of-the-art image understanding algorithms and local feature extraction and matching, such as SIFT, MSER, HOG, VLAD, Fisher Vector, k-means, SLIC superpixels, and others. Written in C for efficiency, with MATLAB interfaces, it’s designed for researchers, students, and developers who need robust and fast tools for image processing, feature extraction, and visual recognition tasks. VLFeat supports Windows, Mac OS X, and Linux, provides comprehensive documentation, and includes applications for classification and image matching as well as command-line and programmatic APIs.