An Improved Method Based on CV and Snake Model for Ultrasound Image Segmentation

In recent years, active contour models have become one of hotspots in the field of image segmentation. However, for ultrasound images, it is difficult to achieve ideal segmentation results with those methods that only based on boundary information or regional information. In this paper, we combined GVF-Snake with CV model. First, preliminary segmentation is implemented with CV model. Then, the local information in each object area is optimized with GVF-Snake. The experimental results demonstrate the proposed method can effectively deal with image blur and edge break in ultrasound images. The resulting contours are relatively intact and close to the manual results.