ABSTRACT
This chapter deals with the subject of utilizing artificial intelligence (AI) for cutaneous imaging. In 2016, researchers published an examination of the validity of using machine learning to generate a quantitative image analysis. For the AI-assisted method to become a viable day-to-day melanoma diagnosis technique, the society needs to resolve several important issues: patients’ trust of this method, medical-legal issues, and proper health insurance coverage. The patients’ trust, in turn, depends on solid clinical data support. Such supportive clinical data would also increase the confidence of physicians, who will be in the interface between patients they care for and the AI applications. Using a deep convolutional neural network (CNN) type of AI, another group of researchers demonstrated that AI could achieve skin cancer classification at board-certified dermatologists’ level of competence. The chapter concludes that CNN could outperform most dermatologists in diagnosing melanoma on dermoscopic images.
