Table 7.
The non-visual, target-focused, and multimodal datasets.
| Dataset | Main Feature | Capacity | Number of Emotions | Environment | Link |
|---|---|---|---|---|---|
| UTSC-NVIE (target-focused) |
The biggest thermal-infrared dataset | Videos of posed and spontaneous expressions | Six | Lab controlled | http://nvie.ustc.edu.cn |
| SPOS (target-focused) |
Subjects with different accessories | 231 images of spontaneous, and posed expressions | Six | Lab controlled | https://www.oulu.fi/cmvs/node/41317 |
| VAMGS (non-visual) |
Visual-audio real world dataset | 1867 images of spontaneous expression 1018 emotional voices |
Six | Real-world | http://emotion-research.net |
| MMSE | A multimodal dataset | 10 GB data per subject | Ten | Lab controlled | http://www.cs.binghamton.edu/~lijun/Research/3DFE/3DFE_Analysis.html |