基于视觉感知机制的轮廓检测方法

Visual perception-based contour detection method

结合初级视皮层的结构及其功能特性,模拟初级皮层的轮廓感知功能,提出一种有效的基于人类视觉感知机制的轮廓检测算法.首先,利用局部能量建立复杂细胞响应模型;然后,结合初级视皮层细胞排列特性构建图像信息与皮层细胞的映射关系,通过模拟神经元间的水平连接机制,实现对图像轮廓信息的增强和背景信息的抑制,建立了以初级视皮层风车状结构为基础的轮廓感知模型.实验结果表明:所提方法在保持目标轮廓完整性的同时,对背景信息的抑制效果更为有效;可以更加有效地对图像背景纹理进行抑制和对目标轮廓进行增强,在一定程度上达到了轮廓过度检测与欠检测的平衡,检测出的目标轮廓位置也更加精准.

An effective method based on human visual perception mechanism was proposed, the structure properties of the primary visual cortex were considered to simulate its function of contour perception. Firstly, the local energy to create the response model of complex cells was used. Then, the mapping relationship between stimulus and neurons in primary cortex with the properties of cells arrangement in it was realized. Finally a contour perception model based on the pinwheel structure of the primery cortex was established. Experimental results show that the proposed method can maintain the integrity of the target outline, meanwhile, the inhibition effect of the background information is more effective; the proposed method can availably enhance the targetrs contour while suppress the cluttered background to achieve a balance between over and lose detection, and has better accuracy of the contour location.