A Flutter prototype that uses AI-powered pose detection to track and count exercise repetitions for multiple exercise types. The earned repetitions can be used to unlock app usage time, promoting fitness and productivity.
- Multi-Exercise Support: Track push-ups, squats, sit-ups (beta), pull-ups, and jumping jacks
- Real-time Pose Detection: Uses Google ML Kit for accurate pose tracking
- KNN Classification: Machine learning-based exercise classification
- Adaptive Thresholds: Calibrate sensitivity for different fitness levels
- Flutter SDK (3.35.0 or later)
- Android Studio
- Dart (>=3.9.2 <4.0.0)
- Clone the repository
git clone https://github.com/BPerju/getfit.git
cd getfit- Install dependencies
flutter pub get- Run the app
flutter run- Define configuration in models/exercise_config.dart
- Create classifier in
classifiers/yourexercise_classifier.dart - Update factory in classifiers/exercise_classifier_factory.dart
- Add CSV asset to pubspec.yaml
- Place training data in
assets/directory
See Original Dataset for reference.
The features were extracted into csv files using the feature extraction script in /feature_extraction/.
- Google ML Kit team for pose detection
- MediaPipe for inspiration on repetition counting
Note: This app requires CSV training data for each exercise. The push-up model is included as a reference. You'll need to collect and format training data for the other exercises following the same structure.