Pictunes - Visual Mood Guided Music Recommendation System
Nelavetla Vaishnavi Reddy, Geeda Dilip Reddy, Vegesna Vidya Sri, Kumbham Deepthi Sri, Subhranginee Das, Pavan Kumar Pagadala · 2024
This paper introduces a novel approach to song recommendations, moving away from conventional methods relying on historical data. Instead, we propose determining a user’s mood from their uploaded picture, utilizing advanced machine learning. By predicting emotional states in real-time, the system offers song recommendations tailored to the user’s current mood. By integrating computer vision and machine learning, this system analyses real-time facial expressions through a device’s camera to infer the user’s emotional state, which could include emotions like happiness, sadness, excitement, or relaxation. A database of music, tagged with emotional characteristics, is used to match the user’s detected emotion with appropriate music tracks, enhancing user engagement and satisfaction. Historical listening preferences are also considered to fine-tune recommendations. This innovative fusion of visual cues and music preferences redefines the recommendation landscape, establishing a more immediate and personalized connection between the user’s emotions and the suggested songs. This approach surpasses traditional reliance on past preferences, introducing a new era of music discovery that aligns seamlessly with individuals’ present emotional states. This approach enhances the user experience by providing music that aligns with their current emotional state, thus improving mood and deepening the connection with the music. This system represents a novel and emotionally intelligent dimension in music discovery, promising a more immersive and satisfying music listening experience, and opening up new possibilities for the evolution of music recommendation systems.