论文部分内容阅读
提出了一种基于人工神经网络的聋儿语音训练识别的多媒体特征挖掘技术 .构造了邻域三层神经元合作竞争的动态行为神经元模型 ;实验选取了动态递减函数、动态最小覆盖矩阵和动态确定阈值形成的 SOL A挖掘算法 ;以及聚类分布的网络图技术 .解决了聋儿语音训练系统中基本语音识别的难题 .
This paper proposes a multimedia feature mining technology based on artificial neural network for speech training recognition of deaf children.This paper constructs a dynamic behavioral neuron model of cooperation and competition among three layers of neurons in the neighborhood.Dynamic diminishing functions, dynamic minimum coverage matrix and dynamic SOL A mining algorithm to determine the threshold value, and network technology of cluster distribution to solve the problem of basic speech recognition in speech training system for deaf children.