QUADTREE DECOMPOSITION COMBINED WITH EDGE DETECTION ENHANCEMENT AND CROSS-CORRELATION FOR OBJECT RECOGNITION AND RETRIEVAL

Abstract

The proposed approach in this paper is region based segmentation technique based on Quadtree decomposition and edge detection. The technique allows an efficient recognition and segmentation of objects in images. Edge detection enhancement method is proposed to overcome some of the most prominent problem associated with this technique. Edges were detected, enhanced, and regions were segmented. Furthermore, cross-correlation between images is proposed as a method to measure the similarity which means that the proposed technique may be useful to be used in CBIR (content-based image retrieval).