Head pose estimation on BIWI Kinect Headpose Dataset (8-fold cross-val)
3.12Yaw ErrorDMLIR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DMLIRProtocol=8-fold cross-validation2018.12 | 3.12 | 4.68 | 3.07 | 3.62 | |
| DMLIREvaluation Protocol=8-fold cross-validation2019.01 | 3.12 | 4.68 | 3.07 | 3.62 | |
| Hybrid coarse-fine classification frameworkEvaluation Protocol=8-fold cross-validation, Backbone=ResNet-502019.01 | 3.4273 | 2.6437 | 2.9811 | 3.0174 | |
| CNN + HeatmapsProtocol=8-fold cross-validation, Input=Uncertainty heatmaps2018.12 | 3.46 | 3.49 | 2.74 | 3.23 | |
| CNN + HeatmapEvaluation Protocol=8-fold cross-validation2019.01 | 3.46 | 3.49 | 2.74 | 3.23 | |
| MLP with locationProtocol=8-fold cross-validation, Input=5 facial keypoints2018.12 | 3.64 | 4.42 | 3.19 | 3.75 | |
| MLP + LocationEvaluation Protocol=8-fold cross-validation2019.01 | 3.64 | 4.42 | 3.19 | 3.75 | |
| DrouardProtocol=8-fold cross-validation2018.12 | 4.24 | 5.43 | 4.13 | 4.6 | |
| DrounardEvaluation Protocol=8-fold cross-validation2019.01 | 4.24 | 5.43 | 4.13 | 4.6 | |
| Ruiz et al.Protocol=8-fold cross-validation2018.12 | 4.81 | 6.606 | 3.269 | 4.895 | |
| Ruiz et al.Evaluation Protocol=8-fold cross-validation2019.01 | 4.81 | 6.606 | 3.269 | 4.895 | |
| LiuProtocol=8-fold cross-validation2018.12 | 6 | 6.1 | 5.7 | 5.94 | |
| Liu et al.Evaluation Protocol=8-fold cross-validation2019.01 | 6 | 6.1 | 5.7 | 5.94 |