A Hybrid Model for Music Recommendation Based on Facial Emotion Recognition
Uuhasri Madala, Soumya Puvvada, Krishna Praneetha Lingamarla, Jaya Sri Annam, Sourav Mondal, Debnarayan Khatua · 2024
In the rapidly evolving landscape of digital music consumption, personalized music recommendation systems have become indispensable tools for enhancing user engagement and satisfaction. However, while current approaches excel in analyzing explicit user preferences and musical attributes, they often overlook the crucial role of emotions in shaping musical experiences. This research proposes a groundbreaking approach to music recommendation based on facial emotion recognition. The dataset used in this study comprises approximately 35,000 images, categorized into seven classes representing different emotion classes. Each image in the dataset is labeled with its corresponding emotion class, enabling supervised learning techniques to be applied effectively.By harnessing the power of facial expressions as indicators of emotional states, our method seeks to offer personalized music recommendations tailored to the user’s current mood and emotional context. Through the integration of advanced techniques in facial emotion recognition and machine learning algorithms, specifically the utilization of a new hybrid methodology AlexNet and Self-Attention-FCN, this paper aims to revolutionize the way music is recommended, emphasizing the importance of emotional resonance in enhancing user experience.