Human Activity Recognition Engineering
As the framework is composed of different building blocks, it contains different folders for each task. The code is structured as follows:
- ml-pipeline
Code to reproduce the experiments for a single modality (IMU data)
- multimodal
Code to reproduce the experiments from the two data sources IMU & Pose Estimations
- tflite-export
Code to reproduce the on-device experiments
- xsens_android
Android code for real-time human activity recognition using the Xsens DOT sensors