Ascites Off Cancer Cell Microscopic Image Classification And Recognition

In these years, incidence of cancer is rising gradually. Lots of information have shown that: once the cancers are diagnosed, 80 percent of them have belonged to the advanced stage and most patients have lost the chance to be cured. So it has become a key problem to detect and cure the cancers in their early stage.At present, the techniques of digital image processing, pattern recognition and artificial intelligence have been more and more used in the biomedicine fields, and has got a lot of delightful fruit. But, among them, there has little work been done on medical image processing used to diagnose the cancer in the early stage. So, according to the modern calculating technology and the practical experience of the pathology experts, and using medical image processing technique to recognize the cancer cells have realistic sense and a bright prospect in medical scientific research and teaching, as well as clinic diagnose.In this paper, based on lots of researching of the present technology fruits, the microscopic images of peritoneal effusion the cancer cells fallen into are processed, and features are extract from them. Then use the BP neural network, the least distance arithmetic and the Bayes classifier to classify and recognize the cancer-cell images. At last, recognizing results of these three methods are compared. Experiment results have proved that the arithmetic and the recognizing results in this paper are sound, and it is a useful and practical tool in computer-aided diagnose.