摘要： 本文提出了一种应用神经网识别低限制的手写数字的新算法,该算法改进了海明网(Hamming net)中匹配测度的计算方法,克服了海明网只能适用于两值模式的限制,使识别率大幅度提高。文中叙述了新算法的原理并给出了实现方法和实验结果。
Abstract: The algorithm presented here recognizes low restrictive handwritten numerals by means of neural networks. It refines the calculating method of the matching score in the hamming net, removes the limitation that the hamming net suits only binary patterns, so that the rate of recognition is increased obviously. The principles of the new algorithm and the results of experiment are given.
林慎机. 基于人工神经网络的低限制手写数字识别[J]. , 1990, 16(06): 3-8.
Lin Shenji. LOW RESTRICTIVE HANDWRITTEN NUMERALS' RECOGNITION BASED ON ARTIFICIAL NEURAL NETS[J]. , 1990, 16(06): 3-8.