Detecting interpretable and accurate scale-invariant keypoints | IEEE Conference Publication | IEEE Xplore