Related Experiment Video
Updated: Jun 1, 2025
07:30
A Test Bed to Examine Helmet Fit and Retention and Biomechanical Measures of Head and Neck Injury in Simulated Impact
Published on: September 21, 2017
8.9K
Can a novel computer vision-based framework detect head-on-head impacts during a rugby league tackle?
Manish Mohan1, Dan Weaving2,3,4, Andrew J Gardner2,5
1Carnegie Applied Rugby Research (CARR) centre, Carnegie School of Sport, Leeds Beckett University, Leeds, UK m.mohan@leedsbeckett.ac.uk.
Summary
A new computer vision framework can automatically detect head-on-head impacts in rugby, aiding in concussion prevention. This technology can analyze video footage to identify risky collisions and inform safety strategies.
Area of Science:
- Sports Science
- Biomechanical Engineering
- Computer Vision
Background:
- Head-on-head impacts are a significant risk factor for concussions in sports.
- Automated detection of these impacts is crucial for player safety.
- Previous research has not evaluated computer vision for head-on-head impacts in rugby.
Purpose of the Study:
- To develop and evaluate a novel computer vision framework for automatically classifying head-on-head impacts in rugby.
- To assess the framework's performance using both standard-definition and high-definition video data.
- To provide a tool for governing bodies to analyze head-on-head collision rates and prevention strategies.
Main Methods:
- A computer vision framework integrating object detection and 3D Convolutional Neural Networks was developed.
- Tackle events from professional rugby league matches were manually coded as head-on-head or non-head-on-head impacts.
- The framework was trained on 341 clips and tested on 670 clips, comparing its classifications against manual coding.
Main Results:
- The framework achieved a sensitivity of 68% and specificity of 84% for head-on-head impact classification in standard-definition video.
- Performance in high-definition video showed similar results with 65% sensitivity and 84% specificity.
- Positive predictive values were 0.61 for both standard-definition and high-definition clips.
Conclusions:
- A novel computer vision framework for detecting head-on-head impacts in rugby has been successfully developed and evaluated.
- The framework can be utilized by sports governing bodies for real-time or retrospective analysis to assess head-on-head collision rates.
- Future research should extend this framework to other head-contact mechanisms and explore its real-time application for clinical assessment support.
