Research On Several Image Segmentation Methods Based On Fuzzy Entropy Theory

Image segmentation is one of the basic techniques of image processing and computer vision. It plays an important role in image analysis and computer vision system. It is a key step of image analysis, comprehension and description. Among all the segmentation techniques, the threshold methods based on probability entropy well consider the randomicity of image signals, but ignore the fuzziness of images, so the segmentation effects are not satisfactory in many cases. As a basic and important issue, image segmentation utilizing fuzzy entropy is intrinsically reasonable and inevitable. Segmentation methods based on fuzzy entropy always segment an image using the threshold value with membership degree 0.5, but they behave worse for images suffering from bad illumination. In this paper, a segmentation method based on generalized fuzzy entropy is proposed. The new method segments an image using the threshold value with membership degree m (0