This paper presents a smart phone based system for storing digital images of skin areas depicting regions of interest (lesions) and performing self-assessment of these skin lesions within these areas. The system consists of a mobile application that can acquire and identify moles in skin images and classify them according their severity into melanoma, nevus and benign lesions. The proposed system includes also a cloud infrastructure exploiting computational and storage resources. This cloud-based architecture provides interoperability and support of various mobile environments as well as flexibility in enhancing the classification model. Initial evaluation results are quite promising and indicate that the application can be used for the task of skin lesions initial assessment.