Magnetic Resonance Images Edge Detection Based on Multi-scale Morphology

Medical image edge detection is the first and the most important steps in the process of extracting geometric features of the objects in medical image process. Conventionally, edge is detected according to some early brought forward algorithms such as gradient-based algorithm, but they are not so good for noise medical image edge detection. A method is proposed to detect the image edge with Gaussian and salt & pepper noise based on multi-scale morphology in this paper. The resource image is classified into two opposite classes by threshold: object and background. The part which is larger than the threshold uses large-scale structure element to detect the edge, and uses small-scale structure element for the part less than the threshold. Large-scale structure element gets through small-scale structure element´s dilation. The experimental results show that the proposed algorithm is more efficient than the usually used gradient-based edge detecting algorithms.