How much of driving is preattentive?
- Submitting institution
-
University of Glasgow
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-11707
- Type
- D - Journal article
- DOI
-
10.1109/TVT.2015.2487826
- Title of journal
- IEEE Transactions on Vehicular Technology
- Article number
- -
- First page
- 5424
- Volume
- 64
- Issue
- 12
- ISSN
- 0018-9545
- Open access status
- Out of scope for open access requirements
- Month of publication
- December
- Year of publication
- 2015
- URL
-
http://eprints.gla.ac.uk/218212/
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
-
- Citation count
- 15
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- ORIGINALITY: Driving from visual input is a complex AI problem. Inattention in drivers is claimed to be a prime cause of road accidents. We use a novel pre-attentive vision-based approach, demonstrating that a driver’s actions can be predicted using only low-resolution peripheral vision, discounting attention. SIGNIFICANCE: This article demonstrates that a cheap predictive model, based on low-resolution video can robustly predict a large proportion of a driver’s actions, making it ideal for embedded applications. Published in a high-impact vehicular technology journal, this extends earlier well-cited papers. RIGOUR: The results are tested on three very different, extensive datasets, confirming the findings.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -