針對用傳統的水平集方法處理心臟MRI（核磁共振圖）這類含有強噪音、弱邊界等
現象的圖像很難取得理想的分割結果的問題，這里將聚類信息和圖像ROI 信息相結合，構造基于新的圖像力的水平集速度函數，該速度函數有較強的抗噪性能，并且能夠處理含有弱邊界、低對比度的圖像。實驗表明該方法具有良好的分割效果。
關鍵詞： 曲線演化；Level Set 方法；MRI；k-均值聚類；圖像力

Level Set method Based on New Image Force Cai Huihua Zhou Yongjin Yang wenguo Zhang Jianwei (Nanjing University of Information & Science Technology, Department of Mathematic, Nanjing Jiangsu 210044, China) Abstract: It is hard to get good segmentation result by using the classic Level Set method for Magnetic Resonance Imaging (MRI) which has strong noises and weak edges. To solve these problems, this paper integrates clustering information of image with information of the region of interesting, and constructs a new velocity function based on this new image force. This new velocity function has better antinoise capability and can deal with images which have strong noises, weak edges and low contrast. The experiments showed the effective of this method.
Keywords: Curve evolving; Level Set method; MRI; k-means Clustering; image force