Models for Emotion Detection and Music Recommendation System using SR-GAN

  • Unique Paper ID: 165802
  • Volume: 11
  • Issue: 1
  • PageNo: 2049-2055
  • Abstract:
  • Emotions are an important part of our life. A living being cannot live without emotions. The emotion of a person is affected by many things in and around it. A person's likes and dislikes vary according to his/her current emotion. Hence if we know the mood of the person we could recommend him/her the products that he would like to have during that situation. In this paper we have proposed a system which predicts the facial emotion of the person and then recommends a song to the person. We have used the FER2013 [9] dataset for this purpose and have used different techniques for data generation, Image Super Resolution, and classification. A Fine Tuned Swin transformer model was used for classification. There were seven classes (emotions) in the dataset. The model could achieve an accuracy of 66.8% on the test-set with a best recall of 0.92 for disgust and happy classes and lowest recall of 0.29 for fear class.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 1
  • PageNo: 2049-2055

Models for Emotion Detection and Music Recommendation System using SR-GAN

Related Articles