仿生模式识别_拓扑模式识别_一种模式识别新模型的理论与应用_王守觉.pdf

10 　　　　　　 Vol.30　No.10 　2002 10 ACTA ELECTRONICA SINICA Oct.　2002 　 ()——— 王守觉 (, 100083) 　　　:　, “”“”. “”, “”, “ ”. , “”.“” ().“”“ ”, . 8800 , 99.75%, 0 0.25%. :　;;; :　O235 　　　:　A　　　:　0372-2112 (2002)10-1417-04 Bionic (Topological )Pattern Recognit ion——— A New Model of Pattern Recognition Theory and Its Applicat ions WANG Shou-jue (Lab of Artif icial Neural Networks , Institute of Semiconductors, CAS , B eji ing 100083, China) Abstract:　A new model of pattern reco nition principles, witch is based on “matter co nition”instead of “matter classification” in traditional statistical pattern reco nition, has been proposed.This new model is better closer to the function of human bein , rather than traditional statistical pattern reco nition usin “optimal seperatin ”as its main principle.So the new model of pattern reco nition is called the Bionic Pattern Reco nition.Its mathematical basis are topolo ical analysis of sample set in the hi h dimensional feature space, therefore it is also called the Topolo ical Pattern Reco nition.The basic idea of this model is based on the fact of the continuity in the feature space of any one of the certain kinds of samples.We did experiments on reco nition of omnidirectionally oriented ri id objects on the same level, with the Bionic Pattern Reco nition usin neural networks, which acts by the method of coverin the hi h di- mensional eometrical distribution of the sample set in the feature space.Many animal and vehicle models(even with rather similar shapes)were reco nized omnidirectionally thousands of times.For total 8800 tests, the correct reco nition rate is 99.75%, the error r