CroCPS: Addressing Photometric Challenges in Self-Supervised Category-Level 6D Object Poses with Cross-Modal Learning

Keyphrases

• 2D Object Detection 25%
• 2D to 3D 25%
• 6D Pose 100%
• Bounding Box 25%
• Category-level 100%
• Category-level Object Pose Estimation 25%
• Challenging Tasks 25%
• Cross-modal Learning 100%
• Cutlery 25%
• Detection Performance 25%
• Domain Gap 25%
• Experiment Results 25%
• Household Objects 25%
• Markerless 25%
• Multimodality 25%
• Polarization Imaging 25%
• Polarization Information 25%
• Pose Annotation 25%
• Real Environment 25%
• Real Image 25%
• Realistic Approach 25%
• Recent Advances 25%
• Reflective Objects 25%
• RGB Image 25%
• RGB-D Data 25%
• Robotic Applications 25%
• Self-supervised Method 25%
• Self-supervision 75%
• Specular Surface 25%
• Synthetic Data 25%
• Synthetic Images 25%

Computer Science

• Annotation 33%
• Detection Performance 33%
• Object Pose 100%
• Pose Estimation 33%
• Robotics Application 33%
• Supervised Method 33%
• Synthetic Datasets 33%