Part heatmap regression (ResNet-152) | 90.7% | Human pose estimation via Convolutional Part Heatmap Regression | - |
Convolutional Pose Machines | 90.5% | Convolutional Pose Machines | - |
Pyramid Residual Modules (PRMs) | 93.9% | Learning Feature Pyramids for Human Pose Estimation | - |
Multi-Context Attention | 92.6% | Multi-Context Attention for Human Pose Estimation | - |
ResNet-152 + intermediate supervision | 90.1% | DeeperCut: A Deeper, Stronger, and Faster Multi-Person Pose Estimation Model | - |
Stacked hourglass + Inception-resnet | 93.9% | Knowledge-Guided Deep Fractal Neural Networks for Human Pose Estimation | - |
Trajectory Space Factorization (F=25) | 83.6 | Trajectory Space Factorization for Deep Video-Based 3D Human Pose Estimation | - |