Melasma Detection Based on Image Processing and Machine Learning

Melasma is a common dermatological condition involving a common form of hyperpigmentation, the dermatological diagnosis of which is plagued by the variable presentation and overlap with other skin conditions. Unfortunately, most of the current diagnostic methods are not very precise or reliable, which results in inadequate patient outcome. According to these challenges, this research develops a Windows based application for detection of melasma using advanced image processing techniques and machine learning algorithms. To detect melasma accurately and efficiently, the proposed system is based on the set of 300 high resolution images from DermNet NZ that is preprocessed, segmented, feature extracted, and classified. The application using C# in Visual Studio achieved 97 % detection accuracy in the application that can be used to enhance patient care and clinical decision making. In this paper, each step of the methodology, system design and results are further described and future directions for the research of melasma detection are provided.