Eye Contact
| Team | Accuracy |
|---|---|
| Ma Fuyan et al. | 0,72619047619047600 |
| Baseline: Gaze + Head Pose (best on val) | 0,57629870129870100 |
| Baseline: Trivial (most likely class on train) | 0,25703463203463200 |
Next Speaker
| Team | Unweighted Average Recall |
|---|---|
| Ma Fuyan et al. | 0,59654038768567300 |
Backchannel Detection
| Team | Accuracy |
|---|---|
| Ma Fuyan et al. | 0,65618619844834600 |
| Garima Sharma et al. | 0,62106982441812900 |
| Baseline: Head + Pose Features (best on val) | 0,59636586361780300 |
| Baseline: All Features | 0,59248672927725600 |
| Baseline: Trivial (most likely class) | 0,50000000000000000 |
Backchannel Agreement
| Team | Mean Squared Error |
|---|---|
| Garima Sharma et al. | 0,06234714713786520 |
| Ma Fuyan et al. | 0,06502104829754260 |
| Baseline: Head Pose Features only (best on val) | 0,06094934530867730 |
| Baseline: All Features | 0,06429582631038260 |
| Baseline: Trivial (mean on train) | 0,06648062283557460 |