基于Hopfield的脱机手写数字识别理论及算法

脱机手写数字识别在很多领域具有广泛的应用前景，国内外学者对此做了大量的研究工作，提出了很多预处理和模式识别的算法，大大提高了手写数字的识别精度。为了提高手写数字识别的精度，文章将Hopfield神经网络应用于脱机手写数字识别分析中，Hopfield神经网络的“能量函数”的能量在网络运行过程中，具有不断地减少最后趋于稳定的平衡状态的特性，而且网络一旦建立即可自动运行，无需要训练。

Off-ling Handwritten Recognition Algorithm based on Hopfield

Off-line handwritten digit recognition has a broad application prospect in many domains, and scholars both at home and abroad have done much research work on this. Many preprocessing algorithms and pattern recognition algorithms are proposed, the accuracy of handwritten digit recognition is thus greatly enhanced. To improve the accuracy of handwritten digit recognition, Hopfield neural network is applied to handwritten digit recognition and analysis. The energy in ＂Energy function＂ of Hopfield neural network has the characteristic of gradually reducing and finally tending to be stable balanced state. Furthermore, once the network is established, it could automatically move without any training.