A Personalized Music Therapy: Emotion-Based Treatment Through User-Centric Emotion Detection

Music serves as a tool for regulating, enhancing, and alleviating unwanted emotional states such as stress, fatigue, or anxiety in daily life. Despite the proliferation of multimedia content and the development of intuitive music players with diverse options, users still find themselves manually sifting through song lists to find ones that match their mood. This report addresses mental health issues by proposing solutions to alleviate stress and anxiety. Additionally, it seeks to streamline the process of finding mood-appropriate music by introducing a highly accurate CNN model for facial emotion recognition. Utilizing a webcam, facial expressions are analyzed to determine emotional states. The CNN classifier, a powerful deep learning model, is utilized to train and test the system for facial emotion recognition using the FER2013 dataset. This dataset, which contains a diverse collection of facial images labeled with various emotions, serves as an excellent benchmark for evaluating the performance of emotion recognition systems. The CNN classifier is implemented using OpenCV, a comprehensive open-source library for computer vision. Through rigorous training and validation, the model achieves an impressive accuracy of 97.42 %. This high level of accuracy indicates that the classifier is highly effective at correctly identifying the emotions depicted in the images. Additionally, the model maintains a low loss value of 0.09, which suggests that the predictions are not only accurate but also consistently reliable across different samples in the dataset.