| Title |
Emotion-aware Contrastive Learning for Facial Expression Manipulation via Mutual Information Perspective |
| DOI |
https://doi.org/10.5573/ieie.2024.61.1.71 |
| Keywords |
Facial expression; Valence and arousal; Emotional intensity; Mutual information |
| Abstract |
Deep learning advancements have expanded facial expression research from recognition to generation. Generating facial expressions from limited information, like emotion classes, is challenging. This study introduces a method to create clearer facial expressions by evaluating mutual information based on emotion intensity. The proposed method's effectiveness is evident with a 28.8 improvement in FID scores and superior pixel reconstruction outcomes in comparative tests. |