Computer aided diagnosis of automatic skin tumour detection for clinical applications

Abstract

Cancer is a serious consequence that arises from the uncontrolled multiplication and change in the structure of cells. These cells divide progressively forming lumps or mass of extra tissues causing a high mortality rate in humans. Skin cancer is one of the most common cancer forms which claims more than millions of precious lives every year in both men and women. According to a survey recorded by the American Institute of Cancer Research, a Melanoma skin tumour is estimated as the 19th most occurring cancer while the non-melanoma is the 5th most occurring cancer. Incident mortality rates are more than twice in men as compared to women. So, early detection of skin abnormalities becomes necessary to help expert radiologists to diagnose and treat patients effectively. Abnormal skin cells are very difficult to diagnose as they have similarities to other normal skin cells, which differ only by their compact nature and the growth of these cells.The key objective of this research work is to provide an accurate identification and an automatic classification of the cancerous skin tumour region from the Computed Tomography (CT) images in terms of region enhancement in terms of accuracy, precision, sensitivity, specificity and response time. The main focus is to estimate the accuracy of sample nodule by Computer-Aided Detection (CAD) in CT images. This novel research work will definitely help the radiologists to identify the malignant or the severe stage of skin cancer and also early-stage or benign nodules to decide the kind of treatment to be delivered to the patients newline