Skin Histopathological Image Used LDED Based Algarithm with Automated Detection of Melanocytes Skin
P. Ramya, V Maderazo Christian · 2014
Skin cancer is the most frequent and malignant type of cancer. Melanoma is the most aggressive type among skin cancers. In melanoma diagnosis, the detection of the melano cytes in the epidermis area is an important step. This paper presents an effective computer-aided technique for segmentation and detection of the melanocytes in the skin histopatho logical image. The nuclei regions in the epidermis area are segmented using the K-means clustering algorithm with k value as 3. K-means clustering algorithm clusters the nuclei region based on space and color information. Then a local region recursive segmentation (LRRS) algorithm is applied to detect the candidate nuclei regions from the initially segmented image. LRRS uses two parameters: intensity and size of nuclei to filter out the candidate nuclei regions. Finally, a novel descriptor, named local double ellipse descriptor (LDED) is applied to differentiate melanocytes from keratinocytes. LDED is based on the double ellipsoidal model and utilizes the local features to distinguish melanocytes from the candidate nuclei regions.