Hybrid neural network for spatio-temporal pattern recognition

A hybrid neural network called TS-LM-SOFM was presented for detecting the spatio-temporal feature. Its top layer is a novel single layer temporal sequence recognizer called TS which can transform sparse temporal sequential pattern into abstract spatial feature representation, while its bottom layer is a modified self-recognizing feature map used as a spatial feature detector. A learning matrix connects the two layers. The experimental results showed that this network can well capture the spatio-temporal feature of input signal.