一个具有拟神经组织的边缘检测模型

AN EDGE DETECTION MODEL WITH NEURON LIKE ORGANIZATION

基于视皮层简单细胞的 Gabor函数模型 ,仿照视皮层的朝向柱结构 ,并辅以模式增强和处理网络 ,构造了一个具有拟神经组织的边缘检测模型 .模拟结果显示 ,其检测能力与平滑 -空间导数型边缘检测算法相当 .分析表明 ,这个基于早期边缘检测器概念的模型与平滑 -空间导数型算法在检测机理上是一致的 .模型为用光学 -电子方法和人工神经网络实现边缘检测提供了一个原型 .

Based on the Gabor function model of simple cortical cells,by imitating the structure of orientation columns in visual cortex, and aided by pattern enhancement and processing nets, a model for edge detection with neuron like organization is constructed. Simulation results show that its detection capability is comparable to that of the advanced edge detection algorithms of the smoothing spacial differentiation type in vogue. It is shown that this model as intuitively based on the early concept of edge detectors is essentially identical in the underlying mechanism to the latter. The proposed model provides a prototype for edge detection using electronic optical techniques and artificial neural nets.