2R-9
Study on segmentation of histopathological images for anomaly detection
○曲　　佳（筑波大），野里博和，坂無英徳（産総研）
The processing of histopathological images has significant impact on the performance of pattern recognition in anomaly detection within histopathological images. In this paper, to improve the efficiency of anomaly detection, we propose a method for segmentation of histopathological images. In this method, according to the color information of stained tissues, we divide the histopathological image into three layers corresponding to nucleus, cytoplasm and background. In the end, we conduct experiments to verify the effectiveness of the proposed method.