In view of domestic cancer medical service diagnosis system to human body cell recognition technology demand, the article studies and designs an effectively based on the pathology analysis human body cell recognition system, and achieved the goal that the accuracy of cancer to diagnose is enhanced.According to the different demand, the article uses the segmentation algorithm of the optimal threshold and sub-edge accumulation algorithm to segmentation of the cell and the cell nucleus. In view of cancer cell's characteristic, extract the cell the basic characteristic, the shape characteristic, the light density characteristic as well as the texture characteristic. The tumor cell type is various and the characteristic is complex, so the article uses FCSN ( fuzzy covering with sphere neighborhood ) algorithm to identifying cancer cells, and it could achieve a good recognition effect. |