A New Method of Medical Image Contour Point Automatic Extraction

摘要

A new method based on mathematical morphology is presented to extract contour points in medical images. Compared with derivative algorithms, the approach is less sensitive to image noise and is more efficient as the thinning operation is not necessary. In this method, we first detect the edge by mathematical morphology to get a clear bilevel image, and then we use the tracing bug to extract contour points. During extracting, we confine the tracing bug by controlling the minimum boundary energy and the degree of the edge curvature. The results show that the algorithm can deal with complex and fuzzy images effectively, and the contour points extracted are in good accordance with the original image.