Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

<i>In vitro</i> transcribed circRNA as a therapeutic agent for cancer.

Acta pharmaceutica Sinica. B·2026
Same author

Serum IgG, definite anti-dsDNA positivity, and advanced HBV-related liver disease: a laboratory-based retrospective study.

Clinica chimica acta; international journal of clinical chemistry·2026
Same author

Advancing proteomic discovery through optimized multi-stage scoring and deep learning-enhanced open search.

Bioinformatics (Oxford, England)·2026
Same author

Molecular characterization and correlation with β-lactam resistance of penicillin-binding protein2x, 2b, and 1a of <i>Streptococcus pneumoniae</i> in clinical pneumococcal isolates.

Microbiology spectrum·2026
Same author

Magnetic Resonance Spectroscopy Deep Learning with Magnetic Resonance Background Generator Enables In Vivo Metabolite Quantification of Hepatic Encephalopathy.

IEEE transactions on bio-medical engineering·2026
Same author

PolyMamba-Net: a lightweight and boundary-aware network for real-time polyp segmentation in colonoscopy.

Frontiers in medicine·2026

Related Experiment Video

Updated: Aug 26, 2025

Tactile Vibrating Toolkit and Driving Simulation Platform for Driving-Related Research
07:15

Tactile Vibrating Toolkit and Driving Simulation Platform for Driving-Related Research

Published on: December 18, 2020

4.6K

Real-time driving risk assessment using deep learning with XGBoost.

Liang Shi1, Chen Qian1, Feng Guo2

  • 1Department of Statistics, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA.

Accident; Analysis and Prevention
|October 3, 2022
PubMed
Summary

This study introduces a deep learning model for real-time traffic crash identification using continuous driving data. The novel approach accurately detects crashes and near-crashes, improving traffic safety management.

Keywords:
Convolutional neural networkCrash predictionDeep learningGated recurrent unitHigh frequency kinematic driving dataNaturalistic driving studyXGBoost

More Related Videos

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.5K
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.0K

Related Experiment Videos

Last Updated: Aug 26, 2025

Tactile Vibrating Toolkit and Driving Simulation Platform for Driving-Related Research
07:15

Tactile Vibrating Toolkit and Driving Simulation Platform for Driving-Related Research

Published on: December 18, 2020

4.6K
A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.5K
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.0K

Area of Science:

  • Deep learning applications in transportation safety.
  • Advanced driver-assistance systems (ADAS) and vehicle dynamics analysis.
  • Data-driven approaches for accident prevention and mitigation.

Background:

  • Traffic crashes are rapid events necessitating real-time prediction for effective safety management.
  • Identifying safety-critical events (SCEs), such as crashes and near-crashes (CNC), is challenging due to their rarity compared to normal driving.
  • High-frequency, high-resolution continuous driving data offers potential for improved crash detection.

Purpose of the Study:

  • To develop a novel deep learning model for accurate and efficient identification of traffic crashes.
  • To leverage continuous driving kinematics data for real-time safety-critical event detection.
  • To address the data imbalance issue inherent in crash and near-crash event identification.

Main Methods:

  • Feature engineering using Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) to capture time-series driving data characteristics.
  • Employing Extreme Gradient Boosting (XGBoost) classifier for high precision and recall in crash identification.
  • Utilizing weighted categorical cross-entropy loss and oversampling techniques to handle imbalanced datasets.

Main Results:

  • The proposed model achieved an overall accuracy of 97.5% in a 3-class classification system (crash, near-crash, normal driving).
  • Precision and recall for crash identification were 84.7% and 71.3%, respectively, outperforming benchmark models.
  • Exceptional recall of 98.0% was achieved for the most severe crash events.

Conclusions:

  • The developed deep learning approach offers an accurate, efficient, and scalable method for identifying traffic crashes from continuous driving data.
  • This model demonstrates significant potential for enhancing real-time traffic safety management and informing the development of safety countermeasures.
  • The findings support the broad application prospects of high-resolution driving data analysis in proactive traffic safety solutions.