Sports ball-tracking computer vision
Prior work · national cricket body (confidential)
Computer vision that follows a small, fast ball across broadcast frames.
Computer vision work for a national sports body, including software that tracks the ball through cricket footage for analysis.
Prior work. A system our engineers delivered in an earlier or concurrent role. The client is confidential, and it isn't presented as a MAUK client project.
- What it is
- AI system
- Industry
- Cross-industry
- Status
- Prototype
- Year
- 2025
Prior work · client confidential
Computer vision
The client's system is not shown
The problem
What needed solving
A cricket ball is small, fast and often motion-blurred, and it has to be followed frame by frame against crowds, pitches and changing light.
The solution
What the team built
Vision software that finds the ball in each frame of footage and joins those sightings into a path analysts can study.
Core functionality
- Finds the ball in every frame
- Rebuilds its path for analysis
Business impact
What it achieved
There's no public figure for this project, and we don't publish numbers we can't back up. What was built is described above.
What sets it apart
Why it works
- 01
Made for a tiny, fast target
Tuned for an object that is small, blurred and moving at speed across busy backgrounds.
Services
Where this work fits
Related projects
Similar projects
ResearchVerifiedRoad-scene object detection on BDD100K
MAUK R&D · computer vision
A road-scene detection model trained from scratch on 70,000 images with a laptop GPU, and an honest account of where it struggles.
0.575detection accuracy (mAP50) on 10,000 unseen images
Machine learning
View project
Demo buildVerifiedCLAW & CO. AI virtual try-on
MAUK showcase store for beauty & nail brands
Shoppers see the set on their own hand before they buy.
AI agents · E-commerce · Machine learning
View project
LiveIn-house buildVerifiedMulti-disease risk screening model
MAUK R&D · applied machine learning
Risk scores for ten conditions from 25 routine measurements, trained on 281,000 records and private enough to run on the patient's own device.
86.88%accuracy on patient records the model never saw
Machine learning
View project
Discuss a similar project
Building something like Ball tracking?
We'll tell you what we'd reuse, what we'd change and what it would take.