Improving 3D human pose estimation via 3D part affinity fields

Keyphrases

• 3D Human Pose 100%
• 3D Human Pose Estimation 100%
• Part Affinity Fields 100%
• 3D Part 100%
• 2D Pose 100%
• Contextual Information 66%
• 3D Pose 66%
• 2D Keypoint 66%
• Neural Network 33%
• Popular 33%
• Computer Vision 33%
• Image pixels 33%
• 2D to 3D 33%
• Three-dimensional Space 33%
• Benchmark Dataset 33%
• Regress 33%
• Existing State 33%
• 3D Mapping 33%
• Two-stage Approach 33%
• Art Performance 33%
• Monocular Image 33%
• Heated Area 33%
• Deep Neural Network 33%
• 3D Coordinates 33%
• Two-stage Architecture 33%
• Simple Regression Model 33%
• Keypoint Detection 33%
• Limb Vectors 33%
• Human3.6M 33%
• Visual Understanding 33%

Computer Science

• Contextual Information 100%
• Pose Estimation 100%
• Art Performance 50%
• Deep Neural Network 50%
• Computer Vision 50%