Facial Recognition Approach using Abcd Algorithm for Cancer Treatment

Abstract

Malignant melanoma is one of the generally known cancers due to the changes in the skin behaviour that cause a drastic increase in numerous melanomas which is seen among many white-skinned people. To detect and classify skin lesions, we require a fast and reliable system. The face detection algorithms are used in which, an image dataset is formed and from that several images are tested for the presence of a face. When the face is present, the image is selected for further processing and separate features are detected. The presence of the face, along with two eyes, nose, mouth and lips are necessary for the face detection to work efficiently. A specific area of the face is selected as a test case and the skin irregularity is checked for abnormal features are present or not. An algorithm by the name Asymmetry, Border, Color and Dermatoscopic features (ABCD) is developed which will check the skin parameters and help figure out the presence of abnormal growth. The accuracy of detection will depend upon the clarity of the input image, the brightness and the sharpness. The later part of the project will stress the importance of data exports from the working data sets to a portable format.