三维异联想Hopfield模型

Three Dimensional Heteroassociative Hopfield Model

三维物体的识别是有重要应用意义的课题 ,神经网络模型是解决该复杂问题的一个途径 .作者将神经网络模型表达为矩阵形式 ,提出异联想矩阵形式的 Hopfield模型 ,从而提出了能较好地解决这一问题的新方案 .与传统Hopfield模型相比较 ,该模型对三维物体的识别有较好的识别率 。

Recognition of three dimensional object is always an interesting subject. Neural network model with its associative capability is an important approach to solving such complex problem. This paper puts forward a heteroassociative Hopfield model in matrix form, which can deal with recognition of three dimensional object with a much smaller interconnection matrix than the original Hopfield model. Thus, this result should be interesting for the development of neural newtork models and for practical applications.