Sportsbox
- #Sports
Sportsbox AI uses computer vision to provide real-time corrective feedback to golf players, helping them to improve their skills and achieve their goals.
- Computer Vision

Impact
The mobile and web-based platform allows:
- 2d/3d golf and baseball models for iOS/Android apps
- 2D Keypoints Detection Model metrics improvements
- WebAPI, optimisation via semi-supervised learning
- Semi-Supervised Learning pipeline implementation
- Created UI for Automatic Data Labeling using existing models
MLOps tasks for model API containerisation in Docker and GCP/Kubernetes Engine deployment
Services we provided
2d/3d golf and baseball models for iOS/Android apps
WebAPI, optimisation via semi-supervised learning
Created UI for Automatic Data Labeling using existing models
Tech Stack
Pytorch
Tensorflow
Pycocotools
Open CV

Challenges and Solutions
🧐 Challenges
- Develop a real-time computer vision algorithm for precise analysis of golf player movements.
- Provide corrective feedback to golf players based on the algorithm’s analysis
💡 Solutions
- Created algorithm which:
- precisely analyses golf player movements
- provides real-time adjustments to optimise swings, posture, and movement patterns
User flow
For Players:
For Coaches: