This chapter presents a mobile application as a decision support system for detecting six types of pigmented skin lesions using deep learning. Melanoma, the most aggressive of these conditions, represents only 1% of skin cancer cases but causes most deaths. Early detection is crucial for a higher chance of cure. In low-income countries, insufficient equipment and specialists make early diagnosis challenging. This mobile application aims to address the problem by allowing nonspecialists to make a probable early detection of skin cancer and thus make the decision to consult a medical specialist. The decision support system uses convolutional neural networks with dense layers and applies the SMOTE [1] method to balance the dataset. Evaluations using the HAM10000 and PAD-UFES databases show a classification accuracy of over 80% across six skin cancer classes, with an improvement of up to 9% when SMOTE is applied. The lightweight application (284 MB) processes images directly from the smartphone camera or stored images, achieving a latency of 9.33 s per response, allowing the processing of six patients per minute. The proposed system offers significant potential to improve early diagnosis and treatment of skin cancer, particularly in resource-limited settings.