Billboard in Focus: Estimating Driver Gaze Duration from a Single Image

2026-01-11

arXiv.org

AI Summary

This study addresses the challenge of automatically assessing the distraction caused by roadside billboards to drivers using only a single image frame, without relying on eye-tracking devices or manual annotations. To this end, the authors propose a fully automated two-stage pipeline: first detecting billboards via YOLO and then predicting gaze duration by integrating their bounding box locations with DINOv2 visual features. Evaluated on the BillboardLamac dataset, the method achieves 68.1% accuracy in gaze duration estimation and attains a billboard detection mAP@50 of 94%. Furthermore, it demonstrates strong generalization capability on Google Street View images. This work represents the first end-to-end, sensor-free approach to quantifying the attentional impact of billboards, offering a novel paradigm for driving safety analysis.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Roadside billboards represent a central element of outdoor advertising, yet their presence may contribute to driver distraction and accident risk. This study introduces a fully automated pipeline for billboard detection and driver gaze duration estimation, aiming to evaluate billboard relevance without reliance on manual annotations or eye-tracking devices. Our pipeline operates in two stages: (1) a YOLO-based object detection model trained on Mapillary Vistas and fine-tuned on BillboardLamac images achieved 94% mAP@50 in the billboard detection task (2) a classifier based on the detected bounding box positions and DINOv2 features. The proposed pipeline enables estimation of billboard driver gaze duration from individual frames. We show that our method is able to achieve 68.1% accuracy on BillboardLamac when considering individual frames. These results are further validated using images collected from Google Street View.

Problem

• driver gaze duration
• billboard detection
• driver distraction
• automated estimation
• roadside advertising

Innovation

• billboard detection
• gaze duration estimation
• YOLO
• DINOv2 features
• driver distraction