Image Segmentation based on The Improved RSF Model

摘要

A new model for image segmentation based on the theory of curve evolution is proposed, which is an improvement of the region-based active contour models. A weighted area term is added into the RSF model, and the coefficient in the area term is changed to a weighted coefficient of robust adaptivity. We use the model to segment medical images. The results show that the new model can reflect the complicated image accurately and its background characteristics. The position of a curve can be fixed by the image information robustly. The adaption of the weighted coefficient of robust adaptivity decreases the mistakes, therefore, get a more efficient image segmentation.