基于神经网络方法的噪声字符识别研究

Research of Noise Words Identification Using Neural Networks

文献[1]提出了一种基于布尔神经网络的字符识别方法.本文采用离散型Hopfield神经网络进行 字符识别.该模型与前者相比在信息处理的并行性和实时性等方面更接近实际生物神经网络.仿真结果表 明,该方法可以有效地对噪声字符进行识别.

A method of identifying noise words using Boolean neural networks is given in Ref.[1]. And another method based on discrete Hopfield neural networks is presented in this paper, where the proposed scheme is more efficient, and closer to the principle of living being's neural networks than the BNN model, in terms of the parallelism and real-time of signal processing. The simulation results show that the latter can be used for recognizing effectively the noise words.