人工神经网络认知器用于模式识别的方法研究

RESEARCH INTO THE METHOD OF PATTERN RECOGNITION BASED ON ARTIFICIAL NEURAL NETWORK PERCEPTRON

认知器是从视觉信息处理机制中导出来的神经网络模型，它在模式识别方面具有独特的能力，适用于视觉信息处理应用。针对认知器的突触连接数量巨大、学习过程计算复杂等问题，通过特殊的训练和识别算法，在微机上模拟认知器神经网络，用它来解决模式识别问题。通过用手写体数字进行训练和识别。

A perceptron is a neural network model which comes from vision information processing system. It has special abilities in pattern recognition, and is suitable to the application of vision information processing. There are some defects in the perceptron. For example, it has a large number of connection points, the calculation of learning process is complex, etc. Concerning the merits and demerits, the perceptron can be simulated in microcomputers through special training and recognition algorithms. The perceptron can be used to solve pattern recognition problems.The results are better by the training and recognition with handwriting digits.