Pattern Recogition Based on Hopfield Neural Network
Zhijun Li · Journal of Wuhan Yejin University of Science and Technology · 2005
The image gathered from the industrial field is often polluted to some degree.To deal with this,a discrete Hopfield neural network,with associative memory characteristics,is applied to the industrial pattern recognition with pollution.The network,characterized by parallel processing,being trainable and fault-tolerant,can recognize and classify the image.In case of input noise pollution,the image will still be discerned and resumed.Polluted industrial images sampled by supervised control system are simulated as examples to prove the validity of the proposed algorithm.