Haiyi Yang ¹ and Ang Li ², ¹ USA, ² California State Polytechnic University, USA

This paper presents a mobile application designed to support mental health through AI-generated art, music, and journaling, guided by emotion detection. The app uses facial emotion recognition and journal sentiment to generate daily personalized images and music that promote emotional healing. Built with Flutter, Python, OpenAI models, and Supabase, the system integrates real-time chat, media storage, and adaptive content generation [1]. Three key systems-emotion detection, AI art, and AI music-are evaluated through experiments and compared with scholarly research. The FER model achieved 70 percent accuracy, and DALL·E-generated images scored highly in emotional alignment [2]. Compared to similar projects, our app stands out for its simplicity, accessibility, and user-driven personalization. Limitations include the need for multimodal input and further validation in clinical settings. Future improvements will focus on enhancing emotional accuracy and user safety. Overall, the project demonstrates a promising and scalable approach to digital art therapy and emotional support.

Mental health, Emotion detection, Art therapy, AI-generated music, Mobile application