Lightweight Spatiotemporal Highway Lane Detection via 3D-ResNet and PINet with ROI-Aware Attention

Dimensions overall score 7.0

GitHub Code Pulse

No public code linked for this paper yet.

• The proposed second model achieves 93.40% accuracy on the TuSimple dataset.result 95%
• The architecture significantly reduces false negatives compared to existing baselines.result 85%
• The proposed approach achieves improved performance with fewer parameters than existing 2D and 3D baselines.technical 90%
• The proposed models are well-suited for integration into Advanced Driver Assistance Systems (ADAS).market 80%
• The second model introduces a Region of Interest (ROI) detection head to selectively focus on lane-relevant regions.method 95%
• The architecture reduces computational complexity through ROI-aware attention.technical 85%
• The first model enhances multi-scale feature representation using Feature Pyramid Network (FPN) and Self-Attention mechanism.method 95%
• The architecture has been validated through real-time inference in the Autonomous Systems Laboratory at City, St George's University of London.technical 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card