Geometry-driven self-supervision for 3D human pose estimation

Keyphrases

• Camera Rotation 100%
• Self-supervision 100%
• 3D Human Pose Estimation 100%
• 3D Annotation 100%
• Computer Vision Algorithms 50%
• Two-module 50%
• Benchmark Dataset 50%
• Superior Performance 50%
• Single Image 50%
• Depth Ambiguity 50%
• Camera Parameters 50%
• Second Module 50%
• Supervised Approach 50%
• Camera Calibration 50%
• 3D Pose 50%
• Fully-supervised 50%
• Self-supervised Method 50%
• 2D Pose 50%
• Pose Alignment 50%
• 3D Posture 50%
• Relative Depth Estimation 50%

Computer Science

• Annotation 100%
• Pose Estimation 100%
• Computer Vision Algorithms 50%
• Depth Estimation 50%
• Superior Performance 50%
• Supervised Method 50%
• Camera Calibration 50%