Detecting human—object interaction with multi-level pairwise feature network

Keyphrases

• Feature Network 100%
• Pairwise Features 100%
• Spatial Configuration 40%
• Visual Features 20%
• Triplet 20%
• Feature Extraction 20%
• State-of-the-art Techniques 20%
• Three-level 20%
• Three-parallel 20%
• Human Action 20%
• Image Understanding 20%
• Human-centric 20%
• Body Parts 20%
• Instance-level 20%
• Parallel Streams 20%
• Three-stream 20%
• Semantic Level 20%
• Semantic Label 20%
• Object Instance 20%
• COCO Dataset 20%

Computer Science

• Object Instance 50%