- Case study ↓
- AI/ML
- Sports Analytics
Padel Analytics That Tracks Every Player and Shot, Automatically
AI that turns padel match video into player tracking, forehand and backhand classification and movement heatmaps, so coaches and players see what really happened on court.
The demo
See Padel Sports Analytics at work
- Industry
- Sports Analytics
- Input
- Padel match video
- Output
- Analytics dashboard, Heatmaps, Annotated video
- Built with
- PyTorch · Detectron2 · I3D · OpenCV · Next.js
The problem
Hours of match video, very few clear answers
Coaches and players record matches, but going through the footage by hand takes hours. Tagging every shot, noting positions and counting rallies is tedious and error-prone, so most video never gets analysed at all.
Without structured data, feedback relies on memory and impressions. It's hard to show a player how they move, which shots they favour or where points are lost.
The solution
What Padel Sports Analytics does
Player trackingEvery player's position and movement through each rally
Shot classificationForehands and backhands labelled automatically instead of tagged by hand
Movement heatmapsCourt coverage mapped with homography, so positions sit on the real court
Analytics dashboardStats and clips in a web portal anyone on the team can use
Annotated match videoReview any rally with tracking and shot labels drawn on
Technology
How it’s built
Every step, from the padel match video to the results.
Step 1
Input
- Court camera
- Recorded match video
Step 2
Detection & tracking
- PyTorch
- Detectron2
- OpenCV
Step 3
Shot classification
- I3D
- TensorFlow
Step 4
Output
- Analytics dashboard
- Heatmaps
- Annotated video
- Detection accuracy on in-game video
- 85%+
- Inference time per frame
- 0.3 s
- Less time spent on manual video analysis
- 70%+
Who it’s for
- Clubs and academiesOffering match analysis to members and players as part of coaching.
- Coaches and playersWho want objective data on movement and shot selection.
- Sports-tech companiesAdding AI match analytics to their own app or platform through an API.
Case study
Padel Sports Analytics: the full case study
- Overview
Axion is a computer vision analytics system for padel players, coaches and sports analysts. It watches match video and automatically tracks player movement, shot types and gestures, giving automatic feedback and performance analysis for training and matches.
- Problem
Analysing padel matches for movement and shot performance by hand is labour-intensive and prone to human error. Traditional systems give no visual insight during play and no detailed breakdown of each player's actions.
- Solution
ML Bench built an AI video analytics system that connects to camera feeds and a web dashboard. An object detection model tracks every player's movement through the game.
An I3D (Inflated 3D ConvNet) model classifies forehand and backhand shots, and homography projection turns player positions into movement heatmaps on the court. Results are shown in an admin dashboard designed in Figma.
- Challenges & solutions
Detection had to be reliable in the target courts, not just on generic sports footage.Trained the models on real-world examples from the target environment.
Analysis had to be fast enough to be useful during training.Optimised the models and ran them on more powerful hardware.
Coaches and players are not technical users.Built a simple web dashboard that anyone can use, with no technical skills needed.
- Results
- 85%+ detection accuracy on in-game video
- Inference time reduced to 0.3 s per frame
- Piloted with 2 sports facilities
- Manual video analysis time reduced by over 70%
Have Questions?
Find the right solution
for now
No. Everything comes from the video, so players play as normal.
Why ML Bench
Why teams choose ML Bench
Built for real-world footageWe build computer vision for real cameras and scans: retail, sports, healthcare, logistics and safety.
Starts with the data you haveExisting cameras, recordings or image archives are usually enough for a first proof of concept.
Custom to your setupWe adapt the model to your camera angles, image quality, classes and workflow, not the other way round.
One team, pilot to productionData, model training, dashboard, API and deployment handled by the same team.
Next AI/ML project →
MRI Segmentation
AI-powered brain MRI segmentation and volumetric analysis for Alzheimer's research, with BIDS standardisation of MRI and PET datasets.
Free assessment
See Your Matches in Numbers
Send us a short clip from your court and we'll show you what the AI picks up, free.
- UI/UX Design
- Web Development
- Mobile Apps
- SEO
- AI/ML Solutions
