Learning 3D-aware Egocentric Spatial-Temporal Interaction via Graph Convolutional Networks

Keyphrases

• 3D Location 20%
• 3D-aware 100%
• Affinity Matrix 20%
• Automated Driving 20%
• Automated Driving Systems 20%
• Behavior Annotation 20%
• Driver Behavior 20%
• Driver Behavior Recognition 20%
• Driving Dataset 20%
• Driving Situation 20%
• Egocentric Interaction 40%
• Extracting Features 20%
• Graph Convolutional Network 100%
• Human-driven 20%
• Interaction Framework 20%
• Interaction Modeling 20%
• Irregular Objects 20%
• Lane Marking 20%
• Pedestrian 20%
• Performance Boost 20%
• Promising Strategies 20%
• Road Users 20%
• Spatial-temporal Interaction 100%
• Traffic Lights 20%

Computer Science

• Affinity Matrix 20%
• Annotation 20%
• Framework Interaction 20%
• Graph Convolutional Network 100%
• Large Data Set 20%