Abstract:
"This MOOD TUNES music recommendation system utilizes facial expressions to determine the user's mood and preferencesDue to its versatility and utility in the present pandemic crisis, this device can function well in both disguised and unmasked circumstances. Although earlier research looked into music suggestion and facial expression recognition separately, the particular difficulty presented by face masks calls for particular consideration. The accuracy of models or the integration of numerous emotions in the context of mask-wearing have not been fully addressed by previous research. This study intends to explore how emotions are perceived and conveyed when facial characteristics like the eyes and brows serve as the main external cues owing to masks that partially
hide the face. The study suggests a unique method for reliably identifying facial expressions of emotion and making music suggestions. Modified activation keys, specific CNN layers, and cutting-edge methodologies are used in this method. In order to properly collect emotional information, the difficulty of face masks is handled by analyzing observable facial characteristics such the eyes and
brows. The suggested approach entails changing CNN architectures with specific layers, attention mechanisms, or feature fusion modules to enhance the extraction of emotional information."
