Medical Image Segmentation Based on an Improved 2D Entropy

Abstract

Medical image segmentation is the basis of medical image three-dimension reconstruction. The accuracy of image segmentation directly affects the results of image 3D reconstruction. Medical image is a kind of grayscale image. In order to adequately utilize gray information and spatial information of image, the traditional 2D gray histogram is improved and forms the 2D D¿value attribute gray histogram. Computation method of average gray and 2D entropy is improved. Use spatial information as a substitute for gray probability to compute entropy. Computation of entropy is based on D-value attribute gray histogram and created spatial different attribute information entropy(SDAIVE). In experiment, a series of head CT images are segmented. Experimental results show that improved threshold method can better segment noise image. This method has strong anti-noise capability and clear segmentation results.