A2J: Anchor-to-Joint Regression Network for 3D Articulated Pose Estimation from a Single Depth Image

ICCV 2019 Fu XiongBoshen ZhangYang XiaoZhiguo CaoTaidong YuJoey Tianyi ZhouJunsong Yuan

For 3D hand and body pose estimation task in depth image, a novel anchor-based approach termed Anchor-to-Joint regression network (A2J) with the end-to-end learning ability is proposed. Within A2J, anchor points able to capture global-local spatial context information are densely set on depth image as local regressors for the joints... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Hand Pose Estimation HANDS 2017 A2J Average 3D Error 8.57 # 1
Pose Estimation HANDS 2017 A2J Average 3D Error 8.57 # 1
Hand Pose Estimation HANDS 2019 A2J Average 3D Error 13.74 # 2
Hand Pose Estimation ICVL Hands A2J Average 3D Error 6.461 # 3
Pose Estimation ITOP front-view A2J Mean mAP 88.0 # 2
Pose Estimation ITOP top-view A2J Mean mAP 80.5 # 2
Hand Pose Estimation NYU Hands A2J Average 3D Error 8.61 # 3