Skip to content

An AI model used to help with the workout session by keeping track of counts and movement.

Notifications You must be signed in to change notification settings

jeronartest/AI-Basketball-Move-Trainer

 
 

Repository files navigation

Basketball Move Trainer

This is a basketball movement trainer that measures the accuracy of the moves that you do and compares them to other common/famous basketball moves to see how accurate you are.

Install OpenCV

Basketball Movements

There are basketball movements that you can measure yourself up against.

Kobe Fade

Running w/ WebCam

To run the kobe fade basketball move using your webcam, go to the command prompt and type in the following command:

python main.py -mt basketball-move -t kobe-fade

Running a Video

To run the kobe fade basketball move using a specific video, go to the command prompt/command line, and type in the following command:

python main.py -mt basketball-move -t kobe-fade -vs {file-path-to-video}

The {file-path-to-video} represents where the video exists on your computer. Here is an example of running the kobe-fade using a pre-set video called Jeron_Fade.mov.

python main.py -mt basketball-move -t kobe-fade -vs Moves/Jeron_Fade.mov

Here is the result of running this video: output_jeron_fade.gif

Accuracy Calculations

The accuracy is measured at specific times that is dependent on the move that the user is trying to perform. Example: for the Kobe Fade movement, it's measured at the release point of the fadeaway. The accuracy is then calculated into a score measuring from 0-100 (0 = not accurate, 100 = highest accuracy). For additional fun, we decided to make different tiers based on the score. Here are the tiers:

  • Him (Score = 80 - 100)
  • Goat (Score = 60 - 80)
  • Legend (Score = 40 - 60)
  • AllStar (Score = 20 - 40)
  • League (Score = 1 - 20)
  • Bum (Score = 0 - 1)

About

An AI model used to help with the workout session by keeping track of counts and movement.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%