Abstract:

The balloon force snake model is sensitive to the location of initial contour and has the shortcoming of weak boundary leakage. An adaptive balloon force snake model based on fuzzy clustering is proposed. This model integrates global information of the image into the balloon force snake model, and is used for image segmentation. It can adaptively change the direction and magnitude of the balloon force according to regional information, which makes the contour converge to the target boundary correctly. Experiment results show that the proposed model has lower restriction on the location of the initial contour and overcome the shortcoming of weak boundary leakage. Therefore, the contour can converge on concave boundary correctly. Furthermore, the improved model has strong robustness to noise and has no critical point problem of Gradient Vector Flow (GVF).

Key words: computer application, image segmentation, Snake model, adaptive balloon force, fuzzy clustering