Abstract: Because of the large volume of medical image data, the impact of noise, blurred boundaries and other reasons, the three-dimensional segmentation process is time-consuming, and easily produces less or over segmentation. To solve the above problems, this paper proposes a three-dimensional fast segmentation algorithm based on Level Set, using Level Set Fast Marching Method to obtain two-dimensional segmental region, optimizing the boundary contour, using the Digital Differential Analyzer method to extract contour pixels, finally introducing the idea of the Scan Line Seed-filling to achieve the three-dimensional fast segmentation. The actual clinical CT images of vertebral segmentation experiment result shows that this method can quickly and accurately separate out the interested area.

Key words: computer application, three-dimensional medical image segmentation, Level Set method, DDA, scan line seed-filling