Binary Robust Invariant Scalable Keypoints (BRISK) was conceived by Leutenegger, Chli, and Siegwart to be an efficient replacement to the state-of-the-art feature detection, description, and matching algorithms. The motivation behind BRISK was to develop a robust algorithm that can reproduce features in a computationally efficient manner. In some cases, BRISK achieves comparable quality of feature matching as SURF, while requiring much less computation time.
The BRISK detector is based on the AGAST detector, which is an extension of a faster performance version of FAST. To achieve scale invariance, BRISK searches for the maxima in a scale space using the FAST score(s) as the comparison parameter. Despite discretizing the scale axis at coarser intervals than in alternative high-performance detectors (for example, the fast Hessian), the BRISK detector estimates the true scale of each keypoint in the continuous scale space...