TRIPOD: The Treadmill walking
IMU, Pedobarographic and Photoelectric
Dataset and Data Processing Pipeline
This is a repository for the publication TRIPOD—A Treadmill Walking Dataset with IMU, Pressure-Distribution and Photoelectric Data for Gait Analysis. The dataset can be found here.
This repository contains a gait analysis data pipeline that can be used to:
- load data from GaitUp Physilog IMUs
- load data from OptoGait
- load data from Zebris FDM-T
- estimate gait parameters from the IMU data
- compare gait parameter of diferent measurement systems
The pipelines components for data loading and gait parameter estimation can be exchanged in order to benchmark different recording systems and algorithms.
The pipeline consists of the following building blocks:
Further information can be found the publication and the documentation.
Clone repository
git clone https://github.com/HPI-CH/TRIPOD.git
cd TRIPOD
Create virtual environment with Python version 3.7 (for example using Conda)
conda create --name name_of_your_choice python=3.7
conda activate name_of_your_choice
Install requirements
pip install -r requirements.txt
Request data for scientific purposes here.
Save data to ./data/raw/
Set configuration in ./src/pipeline_playground.py as needed.
Run python ./src/pipeline_playground.py
This runs the stages of the pipeline for the configured subjects, trials and reference system.
With the default configuration and the sample data provided in the repository, correlation plots and Bland-Altmann plots for three trials of one subject will be created, comparing stride length and stride time between OptoGait and the IMU system.
- Justin Trautmann
- Lin Zhou
The IMU gait analysis algorithms were adapted from Tunca et al. who shared their MATLAB code with us.