【摘要】 提出了递进的基于窄带的多分区C-V方法,并对多幅医学脑部MRI图像进行了分割实验。由于该递进方法分为多个阶段,每阶段只需一个水平集函数,并且在每一阶段应用窄带区域,即只处理窄带区域中的点,从而使计算量大大减少。实验结果证明本算法是有效的,在提高计算速度的同时,可大大改进复杂几何结构的分割效果。更多还原

【Abstract】 A hierarchical multi-phase C-V method based on narrow band was proposed. Several stages are involved in the proposed method, and each stage needs just one level set. Besides, in every segmentation stage only the points of narrow band need to be dealt with. Therefore, the computation work has been greatly reduced compared with the multi-phase C-V method. Experimental results of MRI (Magnetic Resonance Imaging) image show that the proposed method is efficient and stable. The segmentation effect and speed is improved greatly when the image has complex structure.更多还原