Intelligent Infrastructure for Enhancing Vulnerable Road User Safety using Machine Vision Technologies
Bibliographic record
Abstract
Abstract Vulnerable road users (VRUs), such as pedestrians and bicyclists, face a higher risk of severe injuries and fatalities in road collisions, with intersections being particularly hazardous. Enhancing VRU safety at intersections is therefore critical for a safer transportation system. This study introduces a proof-of-concept system capable of detecting VRUs at intersections leveraging image data from vision sensors mounted on roadside infrastructure (e.g., traffic poles). The approach includes the development of a unique VRU detection dataset, comprising labeled images of various VRU types – adults, children, and bicyclists – captured under a range of illumination and weather conditions at real-world public intersections. This dataset addresses a notable gap in VRU detection research, as few datasets offer such environmental diversity from a roadside infrastructure perspective. The dataset was leveraged to train state-of-the-art deep learning models optimized for VRU detection. The models were evaluated using data from both public intersections and a controlled test facility, with particular focus on performance under challenging conditions such as snow and low nighttime visibility. Real-time performance benchmarking of the models was assessed, highlighting their effectiveness in dynamic environments. The results demonstrated that the best model achieved a mean average precision (mAP) of 82% in VRU detection while processing full-HD (1920 $$\times $$ × 1080) frames in real time at 75 ms. Additionally, major challenges in VRU detection at intersections were identified, and recommendations for future research directions were provided.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".