Improvement of Harris Algorithm Based on Gaussian Scale Space

Abstract

Abstract- Features is the description of the image contents which could becorner, blob or edge. Corners are one of the most important feature to describeimage, therefore there are many algorithms to detect corners such as Harris,FAST, SUSAN, etc. Harris is a method for corner detection and it is an efficientand accurate feature detection method. Harris corner detection is rotationinvariant but it isn’t scale invariant. This paper presents an efficient harriscorner detector invariant to scale, this improvement done by using gaussianfunction with different scales. The experimental results illustrate that it is veryuseful to use Gaussian linear equation to deal with harris weakness.