[Morphometric analysis and automatic diagnosis of gastric dysplasia and well-differentiated adenocarcinoma]

A morphometric analysis system for gastric dysplasia and well differentiated adenocarcinoma was set up by the techniques of digital image processing and the theory of mathematical morphology. For 100 cases, seven parameters were measured. A trainable statistic classifier in multidimensional space is designed on the basis of large number of samples obtained from the analysis system. And then we have set up an objective quantitative automatic diagnostic system for gastric dysplasia and well differentiated adenocarcinoma.