An Interactive Method Based on Random Walk for Segmentation of Facial Nerve in NMR Images

摘要

In this paper, a nerves segmentation algorithm is presented based on the combination of hessian matrix and random walk techniques. We used hes-sian matrix to enhance the nerve area, then used morphological operations for nerve skeleton extraction and got the seed for random walk. Our algorithm has three phases, enhancement of nerve area on hessian matrix, extraction of skeleton points on enhanced area, finally segmentation of nerves. The experimental results show that the proposed method works well in segmentation of the nerve in NMR images with high accuracy.