D2D: Divide to Detect, A Scale-Aware Framework for On-Road Object Detection Using IR Camera

Keyphrases

• Image-based 100%
• IR Camera 100%
• RGB Image 100%
• Feature Level 100%
• On-road Object Detection 100%
• Scale-aware 100%
• Inconsistency 50%
• Attention Mechanism 50%
• Object Scale 50%
• IR Image 50%
• Object Detection Network 50%
• Object Detection Dataset 50%
• Scaled-YOLOv4 50%
• Pyramid Architecture 50%
• Multi-level Prediction 50%

Computer Science

• Object Detection 100%
• RGB Image 50%
• Experimental Result 25%
• Attention (Machine Learning) 25%
• YOLOv4 25%