EmoTune: Deep Emotion Detection and Music Recommendation System using MobileNetV3
A. Senthilselvi, S Aakash, Mb. Hariharan, P Abijith · 2023
In today’s interconnected world, prioritizing emotional well-being is essential. This paper introduces "Emo Tune," a music recommendation system elevating users’ emotion through personalized song curation. The system integrates facial recognition, convolutional neural networks (CNNs), and Tamil songs, delivering an immersive user experience. The system excels in precise emotion recognition via facial expression analysis using CNN. It leverages deep Convolutional Neural Network (DCNN) models with Transfer Learning (TL) for adapting pre-trained models to facial emotion recognition. Challenges in traditional facial emotion recognition (FER) methods stem from limited focus on frontal views. Our approach pioneers profile view inclusion, enhancing accuracy by combining Transfer Learning and DCNN depth. Validation across pre-trained DCNN well known model MobileNet, highlights methodology robustness. We use FER 2013 Dataset to train our model using pre trained model MobileNet to get best performance Combining Transfer Learning, deep neural networks, and multi-perspective analysis defines our project. By identifying emotions adeptly, we personalize music journeys, enriching emotional well-being