A novel fuzzy entropy image segmentation approach based on grey relational analysis

The paper presents a new image segmentation method based on fuzzy entropy function and grey relational analysis. In allusion to the sensitivity towards the noise of the traditional fuzzy entropy method, the grey relational degree is introduced to exactly reflect the degree of the pixels belonging to the object or the background. In the proposed method, the gray values of the current pixel and its neighborhood pixels form a comparative sequence, and then the grey relational degree between the comparative sequence and a reference sequence is computed, based on which the membership function of the fuzzy entropy function is modified. Therefore, the membership of the current pixel is determined not only by its own gray value but also by the gray values of its neighborhood pixels, which help to distinguish the noise from the image more accurately. The segmentation experimental results of the Cameraman and the Tire images show the good performance of the proposed method in reducing the noise.