A Hybrid Architecture for Music Recommendation and Control: Integrating Facial Emotion Recognition and Hand Gesture Interaction

For making the user engage in real time we have made a music recommendation system which is been integrated with the facial emotion recognition and hand gesture. We have used Convolution Neural Network for the classification of emotions through the analysis of facial expressions. At the same time, play, pause, skip, and volume control can be performed using natural hand gesture interaction through a Media Pipe based on 21 hand landmark. To ensure reliable performance under varying lighting conditions and changes in user posture we have used several preprocessing techniques. The performance of the system is evaluated using various parameters such as recommendation relevance, gesture recognition accuracy, and end-to-end latency.