全息记忆和Hamming网的新模型及其在汉字手写体识别中的应用

New Models of Holographic Network and Hamming Net and Their Application in Handwritten Chinese Character Recognition

本文提出有注意力转移机制的改进型全息记忆网与快速收敛的Hamming网的新模型。报导了应用网络模型构造的汉字手写体及字符识别系统的实现。实验结果表明,在我们的实验前提下本模型与系统能有效地工作。

We propose two new models, one for Holographic Memory, another of a fast-converging Hamming net, both with an automatic attention moving function. Satisfactory results have been, obtained after the constructing of a handwritten Chinese-character recognition system using these new models.