Case study
Soccer Match Analytics & Insights
- Client: MiniStats
- Industry: Sports & Media
- AI capability: Computer Vision
The challenge
Twenty-two players, one ball, and a camera that keeps moving
Soccer matches are dynamic environments with variable lighting, camera movement, and frequent occlusions. MiniStats required a system that could not only detect players and the ball but also maintain their identities across the entire match, even when players exit and re-enter the frame in low-quality footage where facial recognition is impossible.
What we built
From match video to match events
Detection, tracking, team assignment, then events read from the ball's movement.
- Stage 1
Player and Ball Detection
For player and ball detection we annotated about 10 videos and trained a YOLOv8 model. - Stage 2
Player and Ball Tracking
For player and ball tracking we used NorFair tracker. - Stage 3
Team Identification
We also detect which player belongs to which team. To accomplish this, we experimented with several methods, including using DBSCAN and k-Means for clustering. In our clustering approach, we utilized different features such as color histograms, shirt images (extracting the upper region of the bounding box analytically), and the complete bounding box. We found that the combination of k-means clustering and the complete bounding box image of a person yielded the best results. - Stage 4
Event Detection Logic
For event detection we detected the position of the ball and the person closest to the ball and detected events based on the ball’s motion. For example, if the ball moved from near one person to another person of the same team, then this was considered a pass. Alternatively, if the ball moved from near one person to another person of another team then this was considered as an intercept.
Demo recording: players and the ball tracked through a match, with events derived from ball movement.
The impact
Coaches get numbers instead of impressions
- Players and ball tracked through the match
- Teams assigned automatically
- Passes and intercepts derived from ball motion
The tool automatically derives insights that help players and coaches identify specific areas for improvement. By quantifying metrics like ball possession duration, successful pass rates, and defensive intercepts, MiniStats provides a scientific basis for player development and tactical analysis.
Keep exploring
Related case studies
Real Estate & PropTechComputer VisionAI Property Image Intelligence for Listing Optimization & Compliance
PropTech founders and real estate leaders understand that property photos are the most critical asset in closing deals. PropTexx needed an automated way to enforce strict listing compliance and extract valuable insights from millions of unstructured real estate images without slow manual review.
- Property features tagged automatically
- 50+
- Compliance violation types detected
- 15+
Retail & E-commerceComputer VisionSneaker Visual Retrieval System
AlwaysLegit needed a way for their team at expos to instantly identify shoes from photos and match them against a massive 150K+ SKU database, often in challenging real-world lighting and backgrounds.
- SKUs searchable from a single photo
- 150K+
- Retrieval time per query
- <1s
Health & FitnessComputer VisionX-ray to 3D CT Reconstruction & Knee Alignment Analytics
Traditional CT scans are several times more expensive than X-rays, often delaying critical orthopedic diagnosis. Our client needed a system to reconstruct 3D CT-grade results from standard 2D X-rays to calculate precise bone alignment angles for surgical planning.
2D to 3D reconstruction. Automated angle calculation (CPAK, mHKAA). Enhanced with GANs.
Read the case study
Book a strategy session
Talk to an AI engineer about your project
Tell us what you want to automate. The first call is a 30-minute working session with an engineer, not a sales pitch.
- Send the form, it takes 2 minutes
- We reply within 1 business day, under NDA if you need it
- A 30-minute call to scope feasibility and next steps