Nuclear segmentation for skin cancer diagnosis from histopathological images

Skin cancer is the most frequent and malignant type of cancer. Melanoma is the most aggressive type among skin cancers and if they are detected at an early stage, they can be completely cured. In melanoma diagnosis, the detection of the melanocytes in the epidermis area is an important step. For the detection of melanocytes, use of histopathological images can be used. With the recent advent of whole slide digital scanners, tissue histopathology slides can now be digitized and stored in digital image form. The digitized images are analysed with advanced image segmentation algorithms and features such as intensity and size of the cell nuclei is used to filter out the candidate nuclei regions. This paper deals with Enhancement, Segmentation and Classification in histopathological images of the skin. The proposed method uses CLAHE algorithm for the image enhancement followed by bilateral filtering. The initial segmentation is achieved through Fuzzy C-Means algorithm and a local region recursive algorithm is performed for the final segmentation results. Elliptical derscriptor is used to obtain region ellipticity and local pattern characteristics to distinguish the melanocytes from the candidate nuclei regions.