Invariant Corner Detection Using Steerable Filters and Harris Algorithm

Abstract

This paper proposes a new invariant corner detection algorithm using steerable filters and Harris corner detection. The steerable filters have better orientation selectivity and multi-orientation image decomposition that provide a useful front-end for image-processing and computer vision applications. Corners, in image analysis, are important features for image registration, stereo matching, motion tracking and object recognition. These corners are referred as interest points or key points for image registration. In this paper, we compare the performance of SUSAN, Harris, and propose corner detectors in terms of consistency.