基于神经网络的黑白图形视觉识别机制模拟

物理刺激引起视觉系统的生理反应．经过大脑的信息加工而形成视觉。现有多种模型从不同角度对视觉机制进行模拟。该文利用人工神经网络技术对黑白图形识别的视觉机制加以模拟，构建了一个两层BP(back-propagatbn)神经网络模型。应用所建网络对一幅目标和背景灰度分布曲线有重叠的图像进行了成功识别．对大脑视觉加工机制作了有益的探索。

The Simulation of Vision Mechanism by Means of a Neural Network to Identify Monochrome Image

Physical stimulus could cause physiological reaction in the vision system, and then it was processed by the brain to form vision. There were many models to simulate the vision mechanism. This study formed a two layer BP (back propagation) neural network to simulate the vision mechanism of identifying monochrome image in which the intensity distribution curves of targets and backgrounds were overlapped. The network was a simple model of brain vision mechanism, which could provide some usefulcues of vision information processing.