Deep Learning-Based Road Lane Quality Assessment from Egocentric Dashcam Video

Key points

• Reliable lane markings are important for both human drivers and automated driving systems, but their condition is still commonly assessed through manual inspection.1
• Existing autonomous-driving datasets primarily support lane detection and scene understanding and usually lack maintenance-oriented labels for lane-marking visibility, continuity, or degradation.1
• We propose three lane-quality measures: Contrast Stability Index (CSI), which captures lane-background contrast using Laplacian responses; Line Continuity Ratio (LCR), which estimates structural continuity using edge and Hough-line evidence; and Regression Lane Confidence Score (RLCS), which measures agreement between predicted and map-derived lane counts on matched frames.1
• In a pilot validation using 50 human-ranked images, these proposed lane-specific metrics show stronger rank agreement with human lane-quality ordering (e.g., Spearman's $ρ=0.7316$ and Kendall's $τ=0.5472$ for LCR) than general image-quality measures such as PSNR, SSIM, BRISQUE, NIQE, and LPIPS.1

Sources (1)

• [1]Deep Learning-Based Road Lane Quality Assessment from Egocentric Dashcam Video
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 04:46 PM

  In this study, we present a geo-referenced smartphone dashcam dataset and baseline framework for lane-marking quality assessment.

  Reliable lane markings are important for both human drivers and automated driving systems, but their condition is still commonly assessed through manual inspection.

Extractive summary: sentences quoted from the sources.