Head Pose Estimation On Biwi

评估指标

MAE (trained with BIWI data)

MAE (trained with other data)

评测结果

各个模型在此基准测试上的表现结果

模型名称	MAE (trained with BIWI data)	MAE (trained with other data)	Paper Title	Repository
DDD-Pose	2.80	4.52	A Data-Driven Approach to Improve 3D Head-Pose Estimation	-
6DRepNet360	-	3.39	Towards Robust and Unconstrained Full Range of Rotation Head Pose Estimation	-
hopenet	-	4.89	Fine-Grained Head Pose Estimation Without Keypoints	-
EHPNet	3.43	-	An Effective Deep Network for Head Pose Estimation without Keypoints	-
6DRepNet	2.66	3.47	6D Rotation Representation For Unconstrained Head Pose Estimation	-
Direct Regression	2.54	-	Deep Ordinal Regression with Label Diversity	-
Hybrid Coarse-Fine	3.0174	-	Hybrid coarse-fine classification for head pose estimation	-
WHENet	-	3.81	WHENet: Real-time Fine-Grained Estimation for Wide Range Head Pose	-
MNN	-	3.66	Multi-task head pose estimation in-the-wild	-
LSR	-	3.519	2D Image head pose estimation via latent space regression under occlusion settings	-
DAD-3DNet	3.98	-	DAD-3DHeads: A Large-scale Dense, Accurate and Diverse Dataset for 3D Head Alignment from a Single Image	-
3DDFA	-	19.068	Face Alignment Across Large Poses: A 3D Solution	-
KEPLER	-	13.852	KEPLER: Keypoint and Pose Estimation of Unconstrained Faces by Learning Efficient H-CNN Regressors	-
TRG (w/ 300WLP)	-	2.75	6DoF Head Pose Estimation through Explicit Bidirectional Interaction with Face Geometry	-
FSA-Net (Caps-Fusion)	-	4.00	FSA-Net: Learning Fine-Grained Structure Aggregation for Head Pose Estimation From a Single Image
img2pose	-	3.786	img2pose: Face Alignment and Detection via 6DoF, Face Pose Estimation	-
WHENet-V	-	3.48	WHENet: Real-time Fine-Grained Estimation for Wide Range Head Pose	-
SRHP-Euler	-	4.13	On the representation and methodology for wide and short range head pose estimation	-
FAN (12 points)	-	7.882	How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)	-
PADACO	-	4.13	Deep Head Pose Estimation Using Synthetic Images and Partial Adversarial Domain Adaption for Continuous Label Spaces	-

0 of 29 row(s) selected.