Methods of image smoothing and compression based on logical neural networks are de-scribed.The basic idea of these methods is first to train networks using the patterns sensitive to certain image transformation,then to detect the responses of the networks to the patterns at cor-responding positions of the image in the process of the networks scanning across the image,to im-plement the desired transformation by the simple arithmetical operation of the responses. It is sim-ilar to the procedure of the adaptive pattern recognition in terms of the semi-transparent mapping from the pattern space into the response space instead of the classical mathematieal mapping.These methods,therefore,have the better self-adaptation,robustness,and the ability of massive parallel distributed processing,compared with the algorithm-based methods.In addition to theoretical analysis,typical experimental results are given.
Yang Guoqing;Ge Hongwei. METHODS OF IMAGE PROCESSING BASED ON LOGICAL NEURAL NETWORKS[J]. Acta Aeronautica et Astronautica Sinica, 1994, 15(5): 570-575.
 Pao Y H. Adaptive pattern recognition end neural networks. Addison-Wesley publishing company Inc, 1989: 7-9  Aleksander I. Morton H. An introduction to neural computing. Chapman and Hall: 1990, 70-90  Aleksander I. Veural competing architectures. MIT. 1989: 118-132  Aleksander I. Thomas M V, Bowden P A. WISARI9 a radical step forward in image recognition. Sensor Review. July.1984,120-124  Aleksander 1, Wilson M J P. Adaptive windows for image processing. IEE pr+eedings. Pt. E. Sept 1985, 132 (5):233-245  Pratt W K. Digital image processing. John Wiley and Sons In, 1978; 359-442