Head Pose Estimation on BIWI (test)
1.7Yaw ErrorPOSEidon+
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| POSEidon+Modality=RGB+Depth, Training Data=BIWI (split)2020.05 | 1.7 | 1.6 | 1.7 | 1.6 | — | |
| FSA-Caps-FusionModality=RGB-based, Training Data=BIWI (split)2020.05 | 2.89 | 4.29 | 3.6 | 3.6 | — | |
| QuatNetFull Range=False2022.02 | 2.94 | 5.49 | 4.01 | 4.15 | — | |
| HPEFull Range=False2022.02 | 3.12 | 5.18 | 4.57 | 4.29 | — | |
| VGG16+RNNModality=RGB+Time, Training Data=BIWI (split)2020.05 | 3.14 | 3.48 | 2.6 | 3.07 | — | |
| 6DRepNetEnd to End=false, 3DMM=false2024.07 | 3.23 | 5.32 | 2.78 | 3.78 | — | |
| 6DRepNetFull Range=True2022.02 | 3.24 | 4.48 | 2.68 | 3.47 | — | |
| MFDNetRepresentation=Rot. matrix, Alignment=unaligned, MTCNN detections=Used2024.01 | 3.4 | 4.68 | 2.77 | 3.62 | — | |
| VGGHeadsEnd to End=true, 3DMM=true2024.07 | 3.47 | 5.24 | 2.65 | 3.79 | — | |
| FSA-NetRepresentation=Euler, Alignment=aligned, MTCNN detections=Used2024.01 | 3.59 | 2.9 | 2.27 | 2.92 | 5.36 | |
| WHENet-VModality=RGB-based, Training Data=300W-LP and combined dataset2020.05 | 3.6 | 4.1 | 2.73 | 3.47 | — | |
| MartinModality=RGB+Depth, Training Data=BIWI (split)2020.05 | 3.6 | 2.5 | 2.6 | 2.9 | — | |
| WHENet-VFull Range=False2022.02 | 3.6 | 4.1 | 2.73 | 3.48 | — | |
| WHENet-V2022.10 | 3.6 | 4.1 | 2.73 | 3.48 | — | |
| WHENetRepresentation=Euler, Alignment=unaligned, MTCNN detections=Not used or not stated2024.01 | 3.6 | 4.1 | 2.73 | 3.48 | — | |
| EHPNet2022.10 | 3.68 | 4.03 | 2.57 | 3.43 | — | |
| DAD-3DNetRepresentation=6D, Alignment=unaligned, MTCNN detections=Used2024.01 | 3.79 | 5.24 | 2.92 | 3.98 | — | |
| DAD-3DNet2022.04 | 3.79 | 5.24 | 2.92 | 3.98 | — | |
| DAD-3DNetEnd to End=false, 3DMM=true2024.07 | 3.79 | 5.24 | 2.92 | 3.98 | — | |
| VGG16Modality=RGB-based, Training Data=BIWI (split)2020.05 | 3.91 | 4.03 | 3.03 | 3.66 | — | |
| TokenHPERepresentation=Rot. matrix, Alignment=unaligned, MTCNN detections=Used2024.01 | 3.95 | 4.51 | 2.71 | 3.72 | — | |
| MNNRepresentation=Euler, Alignment=unaligned, MTCNN detections=Not used or not stated2024.01 | 3.98 | 4.61 | 2.39 | 3.66 | — | |
| OpalRepresentation=Euler, Alignment=aligned, MTCNN detections=Used2024.01 | 3.98 | 3.09 | 2.4 | 3.16 | 5.42 | |
| OpalRepresentation=6D, Alignment=aligned, MTCNN detections=Used2024.01 | 3.98 | 3.24 | 2.42 | 3.21 | 5.48 | |
| WHENetModality=RGB-based, Training Data=300W-LP and combined dataset2020.05 | 3.99 | 4.39 | 3.06 | 3.81 | — | |
| WHENetFull Range=Hybrid (✔X)2022.02 | 3.99 | 4.39 | 3.06 | 3.81 | — | |
| WHENet2022.04 | 3.99 | 4.39 | 3.06 | 3.81 | — | |
| EVA-GCN2022.10 | 4.01 | 4.78 | 2.98 | 3.92 | — | |
| QuatNetRepresentation=Quaternion, Alignment=unaligned, MTCNN detections=Not used or not stated2024.01 | 4.01 | 5.49 | 2.93 | 4.14 | — | |
| img2poseRepresentation=Rot. vector, Alignment=aligned, MTCNN detections=Used2024.01 | 4.04 | 3.12 | 3.03 | 3.4 | 6.23 | |
| RetinaFaceEnd to End=true, 3DMM=false2024.07 | 4.07 | 6.42 | 2.97 | 4.49 | — | |
| TriNetFull Range=True2022.02 | 4.11 | 4.76 | 3.05 | 3.97 | — | |
| TriNetRepresentation=Rot. matrix, Alignment=unaligned, MTCNN detections=Used2024.01 | 4.11 | 4.75 | 3.04 | 3.97 | — | |
| SSR-Net-MDModality=RGB-based, Training Data=BIWI (split)2020.05 | 4.24 | 4.35 | 4.19 | 4.26 | — | |
| FSA-NetFull Range=False2022.02 | 4.27 | 4.96 | 2.76 | 4 | — | |
| FSA-Caps-Fusion2022.10 | 4.27 | 4.96 | 2.76 | 4 | — | |
| FSA-NetRepresentation=Euler, Alignment=unaligned, MTCNN detections=Used2024.01 | 4.27 | 4.96 | 2.76 | 4 | 7.64 | |
| SSR-Net-MD2022.10 | 4.49 | 6.31 | 3.61 | 4.65 | — | |
| FDNFull Range=False2022.02 | 4.52 | 4.7 | 2.56 | 3.93 | — | |
| HopeNetRepresentation=Euler, Alignment=aligned, MTCNN detections=Not used or not stated2024.01 | 4.53 | 3.08 | 2.83 | 3.48 | 6.6 | |
| OpalRepresentation=Euler, Alignment=unaligned, MTCNN detections=Used2024.01 | 4.54 | 5.05 | 2.8 | 4.13 | 7.49 | |
| img2poseRepresentation=Rot. vector, Alignment=unaligned, MTCNN detections=Used2024.01 | 4.56 | 3.54 | 3.24 | 3.78 | 7.1 | |
| Img2Pose2022.04 | 4.57 | 3.55 | 3.24 | 3.79 | — | |
| Img2PoseEnd to End=true, 3DMM=false2024.07 | 4.57 | 3.55 | 3.24 | 3.79 | — | |
| OpalRepresentation=6D, Alignment=unaligned, MTCNN detections=Used2024.01 | 4.58 | 4.65 | 2.71 | 3.98 | 7.3 | |
| HopeNet (α = 1)Full Range=False2022.02 | 4.81 | 6.61 | 3.27 | 4.9 | — | |
| HopeNetAlpha value=12022.10 | 4.81 | 6.61 | 3.27 | 4.9 | — | |
| HopeNetRepresentation=Euler, Alignment=unaligned, MTCNN detections=Not used or not stated2024.01 | 4.81 | 6.61 | 3.27 | 4.89 | 9.53 | |
| HopeNet2022.04 | 4.81 | 6.61 | 3.27 | 4.9 | — | |
| HopeNet (α = 2)Full Range=False2022.02 | 5.17 | 6.98 | 3.39 | 5.18 | — | |
| HopeNetAlpha value=22022.10 | 5.17 | 6.98 | 3.39 | 5.18 | — | |
| DeepHeadPoseModality=RGB+Depth, Training Data=BIWI (split)2020.05 | 5.32 | 4.76 | — | — | — | |
| DeepHeadPoseModality=RGB-based, Training Data=BIWI (split)2020.05 | 5.67 | 5.18 | — | — | — | |
| 3DDFA-V22022.04 | 6.8 | 12.08 | 7.54 | 8.81 | — | |
| 3DDFA-V2End to End=false, 3DMM=true2024.07 | 6.8 | 12.08 | 7.54 | 8.81 | — | |
| FAN2022.10 | 8.53 | 7.48 | 7.63 | 7.89 | — | |
| Fanlandmarks=12 points2022.04 | 8.53 | 7.48 | 7.63 | 7.88 | — | |
| KEPLER2022.10 | 8.8 | 17.3 | 16.2 | 13.9 | — | |
| RingNet2022.04 | 8.82 | 5.37 | 7.82 | 7.34 | — | |
| Dliblandmarks=68 points2022.04 | 16.76 | 13.8 | 6.19 | 12.25 | — | |
| Dlib2022.10 | 16.8 | 13.8 | 6.19 | 12.2 | — | |
| 3DDFA2022.04 | 36.18 | 12.25 | 8.78 | 19.07 | — |