Box-Level Tube Tracking and Refinement for Vehicles Anomaly Detection

Wu, Jie; Wang, Xionghui; Xiao, Xuefeng; Wang, Yitong

doi:10.1109/cvprw53098.2021.00464

Cited by 9 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traffic anomaly detection has been advancing both through design of new learning methods and design of new object tracking methods. In the 2021 AI City Challenge, all top-ranking methods made contributions to tracking methods ( 10 – 12 ).…”

Section: Literature Reviewmentioning

confidence: 99%

Transit Safety System Evaluation and Hotspot Identification Empowered by Edge Computing Transit Event Logging System

Ke,

Lutin,

Wang

et al. 2023

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

This paper discusses the importance of near-crash events and associated metadata as valuable sources for smart transit applications, such as surrogate safety measures for transit safety research. The STAR Lab at the University of Washington, sponsored by the Federal Transit Administration, has developed an edge computing system that processes onboard videos for near-crash detection. This paper builds on previous work by addressing two research questions: first, how to leverage the near-crash detection system to synthesize rich data sources on transit vehicles, and second, how to use the smart data hub to support transit operation and safety studies. The proposed procedures for event-based transit data collection, evaluation of commercial collision avoidance warning systems (CAWS) technologies, and transit safety hotspot identification are detailed. CAWS’ performance was benchmarked on four transit buses that were operated for almost a year in Pierce County, WA, U.S. Furthermore, the meta-information of near-crash events enables hotspot analysis and the identification of several exemplar clusters that can be explained by driver behavior and roadway geometries. The results of the experiments demonstrate the system’s promising performance and its applicability to addressing various transit operation questions.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Transit Safety System Evaluation and Hotspot Identification Empowered by Edge Computing Transit Event Logging System

Ke,

Lutin,

Wang

et al. 2023

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

show abstract

“…The existing approaches to vehicular anomaly detection mainly fall into two categories. The first set of approaches [4,9,26,41,52] focus on detecting severe events that cause vehicles to stop, and turn the problem into detecting stalled cars via computer vision from surveillance videos. The second set of approaches [3,16,31,34] focus on single-car abnormal driving behaviors, such as speeding and abrupt braking.…”

Section: Introduction 1motivation and Challengesmentioning

confidence: 99%

Detecting Socially Abnormal Highway Driving Behaviors via Recurrent Graph Attention Networks

Zhang

Wang

et al. 2023

Proceedings of the ACM Web Conference 2023

View full text Add to dashboard Cite

With the rapid development of Internet of Things technologies, the next generation traffic monitoring infrastructures are connected via the web, to aid traffic data collection and intelligent traffic management. One of the most important tasks in traffic is anomaly detection, since abnormal drivers can reduce traffic efficiency and cause safety issues. This work focuses on detecting abnormal driving behaviors from trajectories produced by highway video surveillance systems. Most of the current abnormal driving behavior detection methods focus on a limited category of abnormal behaviors that deal with a single vehicle without considering vehicular interactions. In this work, we consider the problem of detecting a variety of socially abnormal driving behaviors, i.e., behaviors that do not conform to the behavior of other nearby drivers. This task is complicated by the variety of vehicular interactions and the spatial-temporal varying nature of highway traffic. To solve this problem, we propose an autoencoder with a Recurrent Graph Attention Network that can capture the highway driving behaviors contextualized on the surrounding cars, and detect anomalies that deviate from learned patterns. Our model is scalable to large freeways with thousands of cars. Experiments on data generated from traffic simulation software show that our model is the only one that can spot the exact vehicle conducting socially abnormal behaviors, among the state-of-the-art anomaly detection models. We further show the performance on real world HighD traffic dataset, where our model detects vehicles that violate the local driving norms. CCS CONCEPTS• Information systems → Spatial-temporal systems; Data mining.

show abstract

LRATD: a lightweight real-time abnormal trajectory detection approach for road traffic surveillance

Zhang

Ren

2022

Neural Comput & Applic

View full text Add to dashboard Cite

Box-Level Tube Tracking and Refinement for Vehicles Anomaly Detection

Cited by 9 publications

References 17 publications

Transit Safety System Evaluation and Hotspot Identification Empowered by Edge Computing Transit Event Logging System

Transit Safety System Evaluation and Hotspot Identification Empowered by Edge Computing Transit Event Logging System

Detecting Socially Abnormal Highway Driving Behaviors via Recurrent Graph Attention Networks

LRATD: a lightweight real-time abnormal trajectory detection approach for road traffic surveillance

Contact Info

Product

Resources

About