The automatic detection algorithm of tongue cancer stem cells based on fuzzy pattern recognition

In this paper, we present a novel recognition algorithm for detecting tongue cancer stem cells with respect to appropriate scaling factors. Our method can be achieved by computer image processing in the condition that the cancer cells are undifferentiated or slightly differentiated, which is of important research significance in the realm of oral medicine. According to the biological natures of tongue cancer stem cells, we select the curvature variance of cell contour, the nuclear-cytoplasmic area ratio, and the average optical density of cytoplasm as the measurement parameters. Using these three biological parameters, the characteristics of cancerous tumor cells can be described and thus classified. Therefore, those cells can be categorized under the principle of maximum degree of membership in fuzzy pattern recognition algorithms. In this way, the tongue cancer stem cells can be automatically detected. Desirable recognition results given by our experiments have substantiated the efficiency of our algorithm.