Skip to content

nisigit/HAR-App

 
 

Repository files navigation

Android application to parse live movement data from Nordic Thingy and Respeck sensors using Bluetooth. The app also classifies 44 stationary and dynamic activities, including running, hyperventilation, coughing etc., with a 97% accuracy. The TensorFlow classification models use Convolutional Neural Networks and were deployed using TFLite.

About

Android application for detecting and recording human activities using accelerometer and gyroscope readings from a Respeck sensor

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Kotlin 74.7%
  • Java 25.3%