Novel Interest Point Detector Using Bilateral-Harris Corner Method

The Interest point detection algorithm plays a vital role in computer vision applications. The most commonly used interest point detector is scale invariant feature transform (SIFT). The SIFT algorithm fails to match interest points on the edge due to Gaussian filter. In order to overcome this failure a bilateral-Harris corner detector has been proposed. Bilateral filter is an edge-preserving, noise removing and causing smoothening of images. In this paper the bilateral-Harris corner detector is proposed to preserves edge by smoothening and removing noise in an image. The Harris corner point is used to extracts stable and reliable interest points. This detector has been simulated for two evaluation criteria such as repeatability and matching score. An Extensive experimental result shows that the proposed interest point detector is more robust to different image transformations.