Classification of breast cancer cells by image method

The incidence rate of breast cancer ranked first among all malignancies in women of China in recent years. A general cytologic mass screening by fine needle suction method is urgent. In this paper, a process of computerized image processing and classification method of cancer cells is described. After the cell images are scanned from microscope, the whole process of filtering, segmentation, background clearing, feature extraction, feature screening, and classification, all are automatically completed by softwares. A correct classification rate of 93.5% is estimated by leave-one-out method. The result shows that, attention paid to nucleus only will be enough