Deep Cross-modal Representation Learning and Distillation for Illumination-invariant Pedestrian Detection

Keyphrases

• Distillation 100%
• Cross-modal Learning 100%
• Illumination Invariance 100%
• Pedestrian Detection 100%
• Thermal Image 75%
• RGB Image 75%
• Learning Module 50%
• Cross-modal Representation 50%
• Pedestrian 25%
• Segmentation-based 25%
• Effective Solutions 25%
• Multiple Layers 25%
• Loss Function 25%
• Multispectral Data 25%
• Multimodal Network 25%
• Robust Performance 25%
• Feature Detection 25%
• Semantic Level 25%
• Knowledge Distillation 25%
• Feature Segmentation 25%
• Multimodal Features 25%
• Module Base 25%
• Detection Network 25%
• Auxiliary Task 25%
• KAIST 25%
• Multi-task Loss 25%
• Distillation Method 25%
• Distillation Loss 25%
• Light Variations 25%
• Cross-modal Knowledge Distillation 25%
• Aggregation Strategy 25%
• Pedestrian Detectors 25%
• Learning Methods 25%
• Latent Relationships 25%

Engineering

• Experimental Result 100%
• Effective Solution 100%
• Semantic Level 100%
• Features Detection 100%
• Multispectral Data 100%
• RGB Image 100%
• Loss Function 100%

Computer Science

• Illumination Invariant 100%
• Representation Learning 100%
• thermal image 100%
• Knowledge Distillation 66%
• Cross-Modal Feature Learning 66%
• Features Detection 33%
• RGB Image 33%
• Effective Solution 33%
• Experimental Result 33%
• Multispectral Data 33%