Learning Spatio-Temporal Features via 3D CNNs to Forecast Time-to-Accident

Taif Anjum, Louis Chirade, Beiyu Lin, Apurva Narayan

2023

Abstract

Globally, traffic accidents are one of the leading causes of death. Collision avoidance systems can play a critical role in preventing accidents or minimizing their severity. Time-to-accident (TTA) is considered the principal parameter for collision avoidance systems allowing for decision-making in traffic, dynamic path planning, and accident mitigation. Despite the importance of TTA, the literature has insufficient research on TTA estimation for traffic scenarios. The majority of recent work focuses on accident anticipation by providing a probabilistic measure of an immediate or future collision. We propose a novel approach of time-to-accident forecasting by predicting the exact time of the accident with a prediction horizon of 3-6 seconds. Leveraging the Spatio-temporal features from traffic accident videos, we can recognize accident and non-accident scenes while forecasting the TTA. Our method is solely image-based, using video data from inexpensive dashboard cameras allowing for an accessible collision avoidance tool that can be integrated with any vehicle. Additionally, we present a regression-based 3D Convolutional Neural Network (CNN) architecture that requires significantly less parameters compared to its counterparts making it feasible for real-time usage. Our best models can estimate TTA with an average prediction error of 0.30s on the Car Crash Dataset (CCD) and 0.79s on the Detection of Traffic Anomalies (DoTA) dataset elucidated by the longer prediction horizon. Our comprehensive experiments suggest that spatio-temporal features from sequential frames perform significantly better than only spatial features extracted from static images.

Paper Citation

in Harvard Style

Anjum T., Chirade L., Lin B. and Narayan A. (2023). Learning Spatio-Temporal Features via 3D CNNs to Forecast Time-to-Accident. In Proceedings of the 15th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART, ISBN 978-989-758-623-1, pages 532-540. DOI: 10.5220/0011697900003393

in Bibtex Style

@conference{icaart23,
author={Taif Anjum and Louis Chirade and Beiyu Lin and Apurva Narayan},
title={Learning Spatio-Temporal Features via 3D CNNs to Forecast Time-to-Accident},
booktitle={Proceedings of the 15th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART,},
year={2023},
pages={532-540},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011697900003393},
isbn={978-989-758-623-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 15th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART,
TI - Learning Spatio-Temporal Features via 3D CNNs to Forecast Time-to-Accident
SN - 978-989-758-623-1
AU - Anjum T.
AU - Chirade L.
AU - Lin B.
AU - Narayan A.
PY - 2023
SP - 532
EP - 540
DO - 10.5220/0011697900003393