REAL-TIME EMOTION RECOGNITION USINGMULTIMODAL AI: A RASPBERRY PI-BASED APPROACH FOR ENHANCED SOCIAL INTERACTION

Abstract

Technology has come a long way in recognizing what we say—but understanding how we feel when we say it? That’s still a challenge. Emotion recognition is a growing field in artificial intelligence, with applications ranging from healthcare to customer service. This project focuses on using a Raspberry Pi to detect emotions through both facial cues and voice signals. What makes this approach different is its focus on real-time performance and affordability, aiming to bring emotional intelligence to smaller, embedded systems. We’ve chosen tools like OpenFace for facial analysis and OpenSMILE for voice features, both of which work well within the limited resources of a Raspberry Pi.Emotion-aware systems are increasingly being used in smart environments, including automated customer support, learning platforms, and virtual assistants. However, most existing systems rely on powerful computers and cloud services, which are not ideal for real-time interaction in resource-constrained settings. Our aim is to shift this capability to a compact, edge-based device that can function independently without depending on heavy external infrastructure. The motivation behind this project also stems from the growing interest in making machines more empathetic—able to sense not just commands but the feelings behind them.Another driving force for our work is the educational and accessibility potential. Emotionally intelligent systems could oneday support individuals with autism, anxiety disorders, or communication difficulties by offering subtle feedback or interaction cues. By building the project on the Raspberry Pi, we open up opportunities for hobbyists, students, and researchers to experiment with emotional AI without needing high-end equipment.

Author

Mr. SAM SURESH J, Mr. SATHYA VARMA K, Mr. SATHYA PRAKASH, Mr. CAVIN S