Abstract
This article proposes an information system, DermaAI, based on machine learning, aimed at improving the accuracy of detecting skin damage from photographs. The study employed image preprocessing and analysis methods, as well as classification models based on neural networks. The system architecture consists of several layers and includes processes for data collection, processing, and result presentation. During model training, labeled datasets and data augmentation techniques were used, and the results were evaluated using accuracy, F1-score, and AUC metrics. In addition, the robustness of the model was analyzed across different skin types and anatomical regions. The obtained results confirmed that the proposed approach demonstrates high accuracy and reliability. This system enables early detection of skin diseases, supports medical decision-making, and can serve as a foundation for applications in digital healthcare systems.