Pattern Recognition By a Scaled Corners Detection

Abstract

In this paper we developed a new approach to extract points descriptor used for pattern recognition with Corner detection approach. We used scales of image, each scale was scaled by a scaling factor, detect the corners in each scale, extract the key points descriptor from these corners, and using these points descriptor as key features of recognition in the Hough Transform to classify the Descriptor to its class. We implemented and analyzed SIFT algorithm, corner detection algorithm, and the proposed approach. The experimental results using MATLAB of a proposed approach gave results of recognition with high accuracy.

Keywords: Pattern Recognition; Corner Detection; SIFT; Hough Transform.