Learning capability of an optical associative memory using TFT-LCDs is experimentally shown. The system is based on a neural network model and it uses the orthogonal recollection method. This system can compensate for the optical ununiformity of each element, and can recoganize 10 chinese charactors. The optical crosstalk between two TFT-LCDs is the most important for learning capability.
抄録全体を表示