Movie Recommendation System Based on Facial Emotion Recognition

This study aims to develop a movie recommendation system based on facial emotion recognition. Traditional recommendation systems often rely on a user's viewing history and preferences, overlooking real-time emotional responses. This paper addresses this gap by integrating facial recognition and emotion analysis to improve recommendation accuracy and user satisfaction. We utilized cameras or viewer reaction videos, along with deep learning models, to capture and analyze real-time facial expressions while users watched movies. The analysis included categories such as joy, surprise, sadness, anger, and fear, providing a comprehensive understanding of the viewer's emotional experience. These emotions were translated into sentiment scores and then used to rate movies. By incorporating these emotion-derived ratings, the recommendation system can generate personalized movie suggestions that reflect not only the viewer's historical preferences but also their immediate emotional reactions. This dynamic approach ensures that recommendations are more engaging and tailored, considering the viewer's real-time emotional state to enhance overall satisfaction and enjoyment.