Skip to content
/ momento Public
forked from junsoo37/momento

Motion correction application using on-device AI.

Notifications You must be signed in to change notification settings

kkuoo7/momento

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

momento

Motion correction application using on-device AI, POSTECH EECE492-01 team project.

Introduction

Application that extracts a user's pose from smartphone camera with on-device AI in realtime and compares it to an expert's pose. And then calculate user's pose accuracy. This app can be applied to exercising alone without personal training and practicing dancing.

Pose estimation runs in on-device environment. However it was difficult to implement the motion scoring algorithm using dart because of lack of mathematical operations. So we implemented the algorithm using python and numpy. And then serve this algorithm using Flask server, so that the app can use this algorithm by calling API.

Check details about whole project in
- poster Eng ver.
- poster KR ver.

App

Check entire app code in momentong directory.
App Full Demo

Motion scoring algorithm

Check motion scoring code in server directory.
The motion scoring algorithms is based on "Efficient Body Motion Quantification and Similarity Evaluation Using 3-D Joints Skeleton Coordinates" in IEEE Transactions on Systems, Man, and Cybernetics: Systems (2021).
We implemented the algorithm in the paper by modifying for our project.

Tech Stack

  • App: Flutter
  • Motion scoring algorithm(server): python, numpy, Flask
  • Pose Estimation model: MoveNet (Failed to apply 3D VIBE model in on-device environment.. But keep trying to apply it)

Team with

About

Motion correction application using on-device AI.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Dart 80.4%
  • Python 14.0%
  • HTML 2.7%
  • Ruby 2.0%
  • Swift 0.6%
  • Kotlin 0.2%
  • Objective-C 0.1%