Medical image segmentation based on the diffusion equation and MRF model

Due to the impact of noise and imaging equipment, medical image may appear the problems of spots, microgroove and grayscale inhomogeneity, which seriously affect the segmentation effect and bring difficulty to the analysis of pathological. This paper uses algorithm based on the diffusion equation and MRF model to overcome above problems. At first we use improved diffusion equation to smooth the image, the improved diffusion equation uses morphological operators instead of the Gauss convolution term of Lin Shi operator, it can not only preserve the contour but significantly reduce the amount of calculation. Then we use MRF model to segment the processed image. The experiment show that our algorithm has a good ability of anti-noise, and to solve the problems caused by grayscale inhomogeneity of medical image. © 2014 by Binary Information Press.