Towards Accurate 3D Human Body Reconstruction from Silhouettes

Keyphrases

• 3D Body Shape 100%
• 3D Human Body Reconstruction 100%
• Highly Accurate 50%
• 2D Image 50%
• Repeatability 50%
• Network Architecture 50%
• Learning Approaches 50%
• Computer Vision System 50%
• Multiple Output 50%
• Per-pixel 50%
• Ablation Study 50%
• Camera Viewpoint 50%
• Multi-task Learning 50%
• Convolutional Neural Network 50%
• Body Circumference 50%
• Supervised Learning Method 50%
• Body Measurements 50%
• Joint Location 50%
• Background Variation 50%
• Pose Variation 50%
• CAESAR 50%
• Silhouette Images 50%
• Pose Angle 50%
• 3D Joint 50%
• Body Volume 50%
• Synthetic Data Generation 50%
• Camera Variations 50%
• Anthropomorphic Measurement 50%

Computer Science

• Learning Approach 100%
• Synthetic Data 50%
• Data Generation 50%
• Network Architecture 50%
• Vision Systems 50%
• Supervised Learning 50%
• Multitask Learning 50%
• Pose Variation 50%
• Convolutional Neural Network 50%
• Computer Vision 50%