This paper describes the application of a visual pattern recognition neural network in a hybrid model based automatic target recognition (ATR) system. This neural itwxk forms the feature extrtion front end af the ATR and is derived from the Neocognitron network first proposed by K. Fukushima. Fx complex taiget recognition, modiflcatiis to the basic Neocognitron network paradigm were required to enhance robustrss against image distortions due to undersampling (aliasing) and poor feature selection during training. The focus of the paper will be on the enhancements, their rationale, and on the use of the network as a self-organizing feature extractim element of an KFR. Results of experiments with the overall A['R system against target imagery will be shown and discussed. 0-81 94-0874-3/92/$4.O0 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/21/2015 Terms of Use: http://spiedl.org/terms SP!E Vol. 1709 Applications of Artificial Neural Networks III (1992)135 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/21/2015 Terms of Use: http://spiedl.org/terms