Mean Shift-Based Edge Detection for Color Image

Edge detection is an important process in low level image processing. With the advent of powerful computers, it is now possible to move to the more computationally intensive realm of color image understanding. There are many benefits in doing so including the increased amount of information for object location and processing. However, many proposed methods for color edge detection are computational expensive and are not very robust to the image noise. In this paper, a new method based on mean shift algorithm to detect edge in color images is presented. The gradient-ascent mean shift localizes edges accurately in the presence of noise and provides a good computational performance, being based on local operators. Experimental results show the effectiveness and robustness of propose method