Performance Comparison of Supervised and Semi-Supervised Models for Emotion Recognition

Labeling and collecting all emotion data can be highly costly. To address this issue, we aim to explore the potential and versatility of semi-supervised learning in the field of emotion recognition by comparing the training loss and validation loss of supervised and semi-supervised learning methods. For this, we will use a dataset labeled with the VAD model, which quantifies emotions in three dimensions.

박찬미 (Chan Mi Park), 김금동(Geum Dong Kim), 이민수(Min Su Lee)