CMU Multimodal Opinion Sentiment and Emotion Intensity (CMU-MOSEI) is the largest dataset of sentence-level sentiment analysis and emotion recognition in online videos. CMU-MOSEI contains over 12 hours of annotated video from over 1000 speakers and 250 topics.
Variants: CMU-MOSEI
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Emotion Classification | ConCluGen | Multi-Task Multi-Modal Self-Supervised Learning for … | 2024-04-16 |
| Facial Expression Recognition | ConCluGen | Multi-Task Multi-Modal Self-Supervised Learning for … | 2024-04-16 |
| Emotion Classification | MARLIN (ViT-L) | MARLIN: Masked Autoencoder for facial … | 2022-11-12 |
| Emotion Classification | MARLIN (ViT-B) | MARLIN: Masked Autoencoder for facial … | 2022-11-12 |
| Emotion Classification | MARLIN (ViT-S) | MARLIN: Masked Autoencoder for facial … | 2022-11-12 |
Recent papers with results on this dataset: