Novel Image Segmentation Algorithm Based on Automatic GVF Snake Model

Abstract

To address issues that GVF Snake model needs to initialize the contour manually so that image segmentation cannot be automatically handled, the effect of segmentation is also related to the initial contour, and its efficiency and accuracy of the model are not ideal, a novel algorithm for image segmentation based on automatic GVF Snake model is proposed in this paper. We introduce the concepts of "flat redundant point" and " fitting outsourcing polygon". In our algorithm the discrete feature edge points set of target image is first obtained by SUSAN corner detection algorithm. Secondly, finding the fitting outsourcing polygon of discrete edge points set, Finally, the GVF Snake curve is evolved by using the fitting outsourcing polygon as the initial contour curve to complete the image segmentation. It is not only proved that the fitting outsourcing polygon as the initial contour curve can converge to the real target image edge by the algorithm in theory, but the experimental results also show that the proposed algorithm can effectively improve the segmentation accuracy, greatly reduce the number of iterations of the model, and improve the efficiency of the algorithm.