April 23, 2025
The GIST Editors' notes
This text has been reviewed based on Science X's editorial course of and insurance policies. Editors have highlighted the next attributes whereas making certain the content material's credibility:
fact-checked
preprint
trusted supply
proofread
Researchers develop AI tech that may monitor pitchers utilizing low-resolution video captured by smartphones
College of Waterloo researchers have developed new synthetic intelligence (AI) expertise that may precisely analyze pitcher efficiency and mechanics utilizing low-resolution video of baseball video games.
The system, developed for the Baltimore Orioles by the Waterloo group, plugs holes in rather more elaborate and costly expertise already put in in most stadiums that host Main League Baseball (MLB), whose groups have more and more tapped into information analytics lately.
These methods, produced by an organization referred to as Hawk-Eye Improvements, use a number of particular cameras in every park to catch gamers in motion, however the information they yield is usually accessible to the house group that owns the stadium these video games are performed in.
So as to add away video games to their analytics operation, in addition to use smartphone video taken by scouts in minor league and school video games, the Orioles requested video and AI specialists at Waterloo for assist about three years in the past.
The result’s a relatively easy system referred to as PitcherNet, which overcomes challenges resembling movement blurring to trace the actions of pitchers on the mound, then yields information on metrics together with pitch velocity and launch level from customary broadcast and smartphone video.
A paper on the challenge, PitcherNet: Powering the Moneyball Evolution in Baseball Video Analytics, was offered on the 2024 IEEF/CVF Convention on Pc Imaginative and prescient and Sample Recognition and is obtainable on the arXiv preprint server.
"The Orioles approached us with an issue as a result of they weren't capable of analyze pose positions and, subsequently, the biomechanics of their pitchers at video games that won’t have entry to high-resolution cameras," stated Dr. John Zelek, a professor of methods design engineering and co-director of the Imaginative and prescient and Picture Processing (VIP) Lab at Waterloo.
"The purpose of our challenge was to attempt to duplicate Hawk-Eye expertise and transcend it by producing comparable output from broadcast video or a smartphone digital camera utilized by a scout sitting someplace within the stands."
To assist prepare AI algorithms on the coronary heart of the expertise, researchers created three-dimensional avatars of pitchers so their actions might be considered from quite a few vantage factors.
Data from video processed by the system is offered to biomechanics analysts for the Orioles, who’ve dedicated to collectively funding the challenge for an additional yr.
That information can be utilized to regulate how pitchers throw the ball to enhance efficiency or keep away from accidents, and assess the longer term success and sturdiness of pitching prospects.
"Present expertise has already improved baseball analytics," stated Jerrin Vivid, a Ph.D. pupil who had a number one function within the challenge. "Because it's restricted to house video games, nonetheless, there’s a actual want for options that work in any setting, particularly for scouting. That's the place our system is available in."
Researchers at the moment are exploring the applying of the underlying thought—AI evaluation of participant poses utilizing customary broadcast and smartphone video—to different skilled sports activities, together with hockey and basketball, along with different elements of baseball, resembling batting.
Extra data: Jerrin Vivid et al, PitcherNet: Powering the Moneyball Evolution in Baseball Video Analytics, arXiv (2024). DOI: 10.48550/arxiv.2405.07407
Journal data: arXiv Offered by College of Waterloo Quotation: Researchers develop AI tech that may monitor pitchers utilizing low-resolution video captured by smartphones (2025, April 23) retrieved 23 April 2025 from https://techxplore.com/information/2025-04-ai-tech-pitchers-resolution-video.html This doc is topic to copyright. Other than any truthful dealing for the aim of personal research or analysis, no half could also be reproduced with out the written permission. The content material is offered for data functions solely.
Discover additional
Going prime shelf with AI to raised monitor hockey information 0 shares
Feedback to editors
