Road Hazard Stimuli: Annotated naturalistic road videos for studying hazard detection and scene perception
DOI: 10.3758/s13428-023-02299-8
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper introduces the "Road Hazard Stimuli," a novel dataset designed to address the lack of empirical data on how vision and cognition support safe driving in naturalistic settings. The authors argue that studying perception during actual driving is unethical and dangerous, while simulations fail to capture the complexity and visual fidelity of real-world hazards. To bridge this gap, they curated a set of 750 crowd-sourced dashcam video clips from YouTube and Reddit, comprising 434 hazard clips and 316 no-hazard clips. These videos were selected to reflect the inherent variability of driving environments, including different traffic conventions, times of day, and hazard types such as pedestrians, vehicles, and animals. The dataset is rigorously annotated to support behavioral research. Each video was spatially and temporally annotated to identify the precise moment of hazard onset and the driver’s first visible response, as well as the location of the hazard. Additionally, the authors conducted a rating study with 48 licensed drivers to assess the perceived hazardousness of 1,356 brief (333 ms) excerpts extracted from the source videos. These excerpts were timed to mimic the time pressure of real driving, occurring immediately before the driver’s response. Participants rated the clips on a continuous scale from safe to dangerous. The study also collected driving history data to examine the relationship between experience and hazard perception. The results demonstrate that the hazardousness ratings span the entire scale and exhibit high inter-rater agreement, indicating that observers consistently perceive the danger levels in these dynamic scenes. Crucially, the ratings were robust to participants' driving history, suggesting that perceived hazardousness is a stable property of the stimuli rather than a subjective judgment heavily influenced by experience. The dataset captures a wide range of immediate hazards requiring evasive maneuvers, with most driver responses occurring within 1.5 seconds of hazard onset. The significance of this work lies in providing a standardized, ecologically valid tool for studying dynamic scene perception and hazard detection. Unlike existing datasets designed for computer vision or large-scale naturalistic driving studies, this resource is tailored for perceptual research, offering both hazardous and safe scenarios matched for environmental conditions. It enables researchers to investigate how drivers extract information from complex, time-limited scenes without the ethical constraints of on-road experiments or the fidelity limitations of simulators. This dataset serves as a valuable resource for understanding the visual and cognitive mechanisms underlying safe driving, with potential applications in traffic safety, driver behavior analysis, and computer vision.
Key finding
The Road Hazard Stimuli dataset provides high-quality, annotated naturalistic video clips with consistent hazardousness ratings across observers, offering a reliable resource for studying hazard detection and scene perception.
Methodology
dataset
Sample size: 48
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via scout_discovery on 2026-05-08.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | partial | scout | — | — | 2 | 2026-05-08 |
| archive | success | unpaywall | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-08 |
| promote | success | — | — | — | 1 | 2026-05-08 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- hazard perception
- hazard perception training
- road complexity
- peripheral attention
- useful field of view
- naturalistic crash near crash
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Methodological Resource: dataset resource, tool software, measurement protocol