https://github.com/zackakil/realtime-ai-rugby-coach

My ambitious project to use a combination of ML, hardware, and cloud tech to capture and conquer the rugby pitch.
https://github.com/zackakil/realtime-ai-rugby-coach

360-video ai google-cloud machine-learning real-time reinforcement-learning rugby sport tensorflow

Last synced: 25 days ago
JSON representation

My ambitious project to use a combination of ML, hardware, and cloud tech to capture and conquer the rugby pitch.

• Host: GitHub
• URL: https://github.com/zackakil/realtime-ai-rugby-coach
• Owner: ZackAkil
• Created: 2022-04-21T21:07:35.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-04-21T21:15:10.000Z (almost 3 years ago)
• Last Synced: 2024-12-06T19:11:04.738Z (2 months ago)
• Topics: 360-video, ai, google-cloud, machine-learning, real-time, reinforcement-learning, rugby, sport, tensorflow
• Homepage:
• Size: 1000 Bytes
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Realtime AI rugby coach

My ambitious project to use a combination of ML, hardware, and cloud tech to capture and conquer the rugby pitch.

## Step 1 : Capture

Using 360 video capture to record the entire pitch.

## Step 2 : Track

Use custom object tracking models to track players.

## Step 3 : Transform

Using the player tracks and some mathmatical things that I need to learn, generate a neat 2D orthoganal projection of the player tracks.

## Step 4 : Analyse

Using the clean 2D player tracks, train models to detect when specific plays are excuted and when score happen.

## Step 5 : Learn the game

Feed the time timeline of data (plays, scores, location on pitch) into a reinforcement learning model to devise a model that can predict which plays have the highest chance of resulting in scores. 

## Step 6 : Take it to the pitch

Have all this run in real-time at the side of the pitch to give instant tactical suggestions.