Color image segmentation using visible color difference and canny edge detector

Abstract

This paper presents an improved version for the JSEG color image segmentation algorithm integrating visible color difference and canny edge detector to measure the boundary of the region. This method resolves the major inherent problem in JSEG caused by varying illumination using visible color difference and canny edge detector with JSEG segmented image. Experiments on natural color of Berkeley Segmentation Dataset are presented which show improved results in comparison with the classical JSEG and Fractal-JSEG algorithm.