Nuclei Segmentation for Breast Cancer Analysis

摘要

Automated detection and segmentation of nuclear structures is critical for classification and grading of breast cancer histopathology. In this paper, we present a methodology for detection and segmentation of structures of interest in digitized histopathology images. The scheme integrates image information from across two different scales: (1) Pixel - level information based on pixel values, (2) Marker based Watershed segmentation. In this paper, we focus on the following problem. First is overlapping of the cells, and to detect the boundary lines between the touching cells in a given image of a tissue sample, capture the deviations in the cell structure and the changes in the cell distribution across the tissue, which are possibly caused by cancer. In sample Histopathological images of Breast cancer the cell distribution is completely different for cancerous and healthy tissue. The first challenge is the noise elimination in the task of determining the focal areas in the image. In case of focusing on properties of nuclei/cell in the image, the second challenge is the nucleus/cell segmentation, this is challenging because of the complex nature of the image scenes and the noise.