An efficient system for detection of mitosis for breast cancer screening

Abstract

Breast cancer is threatening disease for the women around the world and timely detection and treatment is required for
breast cancer patients. The current methods for the detection of breast cancer is expensive and having false error rate. In
this paper, an efficient system for breast cancer detection using histopathological cell images is proposed. The luminance
image is separated from breast cell image in preprocessing stage and further texture feature are extracted from preprocessed
cell image. Principal Component Analysis method is used to classify the features and morphological operations are used to
segment the nuclei region form classified mitosis image. The performance metrics as sensitivity, specificity and accuracy
are evaluated for breast cancer screening to show the performance of proposed method.