MoodSync: Personalized Video Recommendation Based on User Face Emotion

Navigating the vast expanse of video content presents a challenge in discovering personalized, user-specific material. One way to tackle this issue is through a video search feature. However, while using keywords helps, it doesn't always give the best user experience. This abstract explores the use of facial emotion in contrast to watch history for better recommendation of videos to the user. YouTube is a widespread online video-sharing platform where users can upload, share, and view videos. Its recommendation system utilizes algorithms that analyses user behaviors, such as watch history, search patterns, and engagement metrics. Machine learning models, like collective filtering and deep neural networks, are employed to predict what content a user may enjoy. This personalized method enhances user experience, encourages longer engagement, and contributes to the platform's success in delivering applicable and engaging video content. We wanted to know what people liked, so we looked at how they felt when watching videos. With this knowledge, a new method of recommending videos that correspond with their feelings was developed. In real-time, our application reads users' facial expressions and recommends YouTube videos accordingly.