Nowadays, computational technology is given great importance in the health care system to understand the importance of advanced computational technologies. Skin cancer or skin disease (melanoma) has been considered in this chapter. As we know, the detection of skin lesions caused by exposure to UV rays over the human body would be a difficult task for doctors to diagnose in the initial stages due to the low contrast of the affected portion of the body. Early prediction campaigns are expected to diminish the incidence of new instances of melanoma by lessening the populace's openness to sunlight. While beginning phase forecast campaigns have ordinarily been aimed at whole campaigns or the public, regardless of the real dangers of disease among people, most specialists prescribe that melanoma reconnaissance be confined to patients who are in great danger of disease. The test for specialists is the way to characterise a patient's real danger of melanoma since none of the rules, in actuality, throughout the communities offer an approved algorithm through which melanoma risk may be assessed. The main objective of this chapter is to describe the employment of the deep learning (DL) approach to predict melanoma at an early stage. The implemented approach uses a novel hair removal algorithm for preprocessing. The kmeans clustering technique and the CNN architecture are then used to differentiate between normal and abnormal skin lesions. The approach is tested using the ISIC International Skin Imaging Collaboration Archive set, which contains different images of melanoma and non-melanoma.
Keywords: Convolutional Neural Networks, Deep Learning, Hair Removal, Melanoma, Sequential Model, Skin Lesions.
