HS-Pose: Hybrid Scope Feature Extraction for Category-level Object Pose Estimation

Keyphrases

• 3D Graph Convolution 100%
• Category-level Object Pose Estimation 100%
• Complex Shape Object 25%
• Convolutional Layer 25%
• Estimation Task 25%
• Feature Extraction 100%
• Geometric Structure 25%
• Global Information 25%
• Global-local 25%
• Invariance Property 25%
• Large Margin 25%
• Latent Feature 25%
• Local Geometric Feature 25%
• Network Structure 25%
• Object Size 25%
• Point Cloud Data 25%
• Robustness to Noise 25%
• Scale Invariant 25%
• Shape Variation 25%
• Simple Network 25%
• State-of-the-art Techniques 25%
• Structure Information 25%
• Translation Invariant 25%

Computer Science

• Baseline Method 25%
• Convolution Layer 25%
• Feature Extraction 100%
• Global Information 25%
• Graph Convolution 100%
• Network Structure 25%
• Object Pose 100%
• Point Cloud 25%
• Pose Estimation 100%
• Shape Variation 25%
• Translation Invariant 25%

Engineering

• Feature Extraction 100%
• Geometric Feature 33%
• Invariant Property 33%
• Limitations 33%
• Metrics 33%
• Point Cloud 33%
• Pose Estimation 100%
• Simple Network 33%
• State-of-the-Art Method 33%