Synthesizing training images for boosting human 3D pose estimation

Keyphrases

• Training Image 100%
• Human 3D Pose Estimation 100%
• Single Image 66%
• Convolutional Neural Network 66%
• Pose Space 66%
• 3D Pose 66%
• Synthesis Methods 33%
• Fully-automatic 33%
• Direct Estimation 33%
• Bridging the Gap 33%
• Estimation Task 33%
• Superior Performance 33%
• Crowdsourcing 33%
• Challenging Tasks 33%
• Real Image 33%
• Synthetic Images 33%
• Domain Adaptation 33%
• Scalable Approaches 33%
• Clothing Texture 33%
• Synthetic Training Data 33%
• Human Pose 33%
• 2D Pose Estimation 33%
• Synthetic Training Images 33%
• 3D Pose Estimation 33%
• Pose Annotation 33%
• Space Coverage 33%

Computer Science

• Training Image 100%
• Pose Estimation 100%
• Annotation 50%
• Convolutional Neural Network 50%
• Superior Performance 25%
• Training Data 25%
• Synthetic Image 25%
• Domain Adaptation 25%