从人类视觉特性出发,将具有“中心抑制,周边加强”性能的人眼视觉侧抑制现象与交叉视觉皮质模型相结合,形成了一种与人类视觉保持一致的适合图像分割的侧抑制-交叉视觉皮质模型.在此基础上,利用灰度级-邻域平均灰度级二维直方图,将一维卡方散度推广至二维,构造出了基于最小二维卡方散度的分割准则,进而提出一种实时性高的图像分割新算法.实验仿真表明,这种分割算法效果优于原始ICM算法和经典的Otsu算法,同时该准则计算速度远高于基于二维最小交叉熵准则的计算速度.%In the view of human visual feature,the Lateral Inhibition of human vision that has features of inhibiting center and enhancing surroundings is combined with Intersecting Cortical Model,for the sake of forming the Lateral Inhibition-Intersecting Cortical Model which brings into correspondence with human visual.Meanwhile,through the gray level-average gray level histogram,1-D Chi-square divergence is extended to 2-D Chi-square divergence,then the segmentation criterion based on the rule of 2-D Chi-square divergence is structured.Further,a new real-time segmentation algorithm is proposed.At last,simulations validate that this algorithm is not only more simple than 2-D cross-entropy one in computing complexity,but also superior to the classical original ICM and Otsu in the performance of outcomes.
展开▼
