Publication Date: 2023/12/23
Abstract: The exciting rise of big data in recent years has drawn a lot of attention to the interesting realm of deep learning. Convolutional Neural Networks (CNNs), a key component of deep learning, have demonstrated their worth, particularly in the field of facial recognition [3]. This research presents a novel and creative technique that combines CNN-based microexpression detection technology with an autonomous music recommendation system [3] [1]. Our innovative algorithm excels at detecting minor facial microexpressions and then goes above and beyond by selecting music that perfectly matches the emotional states represented by these expressions. Our micro-expression recognition model performs admirably on the FER2013 dataset, with a recognition rate of 62.1% [3]. We use a content-based music recommendation algorithm to extract some song feature vectors after we've deciphered the specific facial emotion. Then we turn to the tried-and-true cosine similarity algorithm to do its thing and recommend some music [3]. But it does not end there. This study isn't only about improving music recommendation systems; it's also about investigating how these systems may assist us manage our emotions [2] [1]. The findings of this study offer a great deal of promise, pointing to interesting prospects for incorporating emotion-aware music recommendation algorithms into numerous facets of our life."
Keywords: Deep Learning, Facial Micro-Expression Recognition, Convolutional Neural Network (CNN), FER2013 Dataset, Music Recommendation Algorithm, Emotion Recognition, Emotion Recognition In Conversation (ERC), Recommender Systems, Music Information Retrieval, Artificial Neural Networks, Multi-Layer Neural Network.
DOI: https://doi.org/10.5281/zenodo.10427665
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23DEC1027.pdf
REFERENCES