• Version
• Download 5
• File Size 218.53 KB
• File Count 1
• Create Date 9 May 2023
• Last Updated 9 May 2023

EMOTION BASED MUSIC RECOMMENDATION SYSTEM

ARPAN SHRIMALI

DEPARTMENT OF COMPUTER SCIENCE

BIRLA INSTITUTE OF TECHNOLOGY

MESRA-835215, RANCHI

Noida Campus

ABSTRACT

Music can make us emotional; for this reason it is often referred to as the "language of thought". This work explores theory in addition to labels of deep learning models to improve existing music.

We've taken a whole new approach to auto playing music  using facial expressions.. Most old or current methods involve playing music manually. We use convolution neural networks for emotion detection.. For music recommendations, Machine learning concepts are used. Our proposed system tends to reduce the calculation and computational in obtaining the results and the overall cost of the designed system, so increasing the system’s overall accuracy. Testing of the system done on dataset FER 2013. Facial expressions are captured by web camera. Emotion detection is done on provided face video to detect emotions such as natural, angry, surprise, happy, sad. By detecting current emotion automatic playlist is generated.

This exercise will collect thoughts on the specific meaning of  each song in the sample file. Because the effect of music on emotions is subjective and varies from person to person, this study will require a large sample size to reduce content. Due to limited resources, some of the knowledge will be taken from the disciplinary perspective and the rest from the active learning model. It uses GAN (Generative Adversarial Network) to create consensus-based content. This work leads to two types of feedback, emotional and non-emotional. Two measures, cosine similarity and Euclidean distance, are used to evaluate the accuracy of the model. The results showed that the difference was not significant and the utility model was not relevant. It can be concluded that consensus as a feature in aesthetic vision is promising. More research is needed to  use better resources to reduce some problems and improve the extraction of emotional information.

Keyword-

Collaborative filtering, Emotion Recognition, Linear classifier, Facial Landmark Extraction, SVM Classification, Learning (Artificial Intelligence), Music, Recommender Systems, Machine Learning