Lesion boundary detection in ultrasound breast images

摘要

In this paper, we proposed a statistical noise estimation kernel based on Rayleigh distribution function. The kernel, as a filter, is implemented in breast lesion boundary detection for ultrasound images. The detection sequence includes: histogram equalization, nonlinear diffusion, convolution of the breast image with the developed kernel, median filtration, segmentation, labelling and morphological operations for the final edge detection. The method used in this work, highlights lesion boundaries effectively.