ABSTRACT
Automatic detection of skin cancer in the dermoscopic image is required for detecting melanoma at an early stage. Skin cancer is extremely dangerous, and it is curable when diagnosed at an early stage. The most common cause of skin lesions is due to the infection of the skin by bacteria, viruses, fungi, or parasites. Previously, skin cancer was diagnosed by biopsy. Later the lesions occurring on the surface of the skin can be detected by analyzing the dermoscopic images of the skin. However, doing manual analysis is time consuming, and it is also difficult to differentiate benign from malignant. Any automated skin detection system can aid in the early detection and also can help in differentiating the tumors. However, attaining high sensitivity and specificity is still challenging due to the unavailability of more malignant samples. In order to cure the human skin cancer with accuracy and also to reduce the error-prone due to human, dermatoscope is in use nowadays. In this review, we discussed the state-of-the-art deep learning (DL) form the foundation of machine learning as it provides better accuracy for medical images.
