Detecting Mental Health Disorder using Deep Learning

Early detection of mental health disorders is crucial for timely intervention and improved well-being. This work presents a web-based application that integrates deep learning techniques to analyze emotions through multimodal inputs: facial expressions and text-based interactions. Using Google FaceNet and Tensor.js, the system performs real-time facial emotion detection (happy, sad, anger) directly in the browser. The Gemini API is used for processing textual inputs via a chatbot powered by an LLM. This allows for meaningful conversations and dynamic questioning to be used in determining emotional states. The front end of the application is built with Next.js for an interactive user interface, and the backend utilizes Python Flask for robust functionality. SQLite is used as the database for efficient data management. These integrated technologies show promise as a widespread solution for mental health surveillance since they confirm AI solutions can help detect disorders at an early stage and protect emotional wellness.