Contrastive learning-based knowledge distillation for RGB-thermal urban scene semantic segmentation

Engineering

• Tasks 100%
• Feature Extraction 100%
• State-of-the-Art Method 100%
• Similarities 100%
• Conducted Experiment 100%
• Receptive Field 100%
• Deep Learning 100%
• Level Feature 100%
• Surrounding Environment 100%
• Semantic Level 100%

Computer Science

• Image Segmentation 100%
• Contrastive Learning 100%
• Knowledge Distillation 100%
• Feature Interaction 66%
• Deep Learning 33%
• Receptive Field 33%
• Feature Extraction 33%
• Autonomous Driving 33%
• segmentation accuracy 33%
• Edge Information 33%
• Parameter Size 33%
• Terminal Device 33%
• Distribution Information 33%
• Surrounding Environment 33%
• Distinctiveness 33%
• Gain Knowledge 33%