Multi-level image segmentation using fuzzy clustering and local membership variations detection

A segmentation method for gray-level images with fuzzy clustering and local detection of membership variations is presented. The method is very efficient for edge detection in images where transitions between two regions are very large. Two fuzzy operations and a fuzzy c-means algorithm adaptation for pixel clustering are introduced. The influence of the number of clusters on the results is discussed. The results obtained by application of the method to noisy and nonnoisy edges are compared, with those obtained by using the gradient operator.<
>