Compositional Human Pose Regression
Xiao Na Sun, Jiaxiang Shang, Shuang Liang et al. · IEEE International Conference on Computer Vision (ICCV) · 2017
Regression based methods are not performing as well as detection based methods for human pose estimation. A central problem is that the structural information in the pose is not well exploited in the previous regression methods. In this work, we propose a structure-aware regression approach. It adopts a reparameterized pose representation using bones instead of joints. It exploits the joint connection structure to define a compositional loss function that encodes the long range interactions in the pose. It is simple, effective, and general for both 2D and 3D pose estimation in a unified setting. Comprehensive evaluation validates the effectiveness of our approach. It significantly advances the state-of-the-art on Human3.6M [20] and is competitive with state-of-the-art results on MPII [3].
Cited by 542 works. doi:10.1109/iccv.2017.284 Open-access PDF
Authors
References
- Deep Residual Learning for Image Recognition
- ImageNet: A large-scale hierarchical image database
- Caffe
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Pictorial Structures for Object Recognition
- Stacked Hourglass Networks for Human Pose Estimation
- Generalized Procrustes Analysis
- 2D Human Pose Estimation: New Benchmark and State of the Art Analysis
- Joint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation