Multi Modal Gait Database

This is the repository to the article
V.Losing & M. Hasenjäger, A Multi-Modal Gait Database of Natural Everyday-Walk in an Urban Environment, 2022. We provide some scripts to generate custom pandas data frames from the raw data to streamline the processing and machine learning. Additionally, there is also a visualization / labeling tool enabling data inspection and the modification of the labels.

Example videos

These videos provide an impression of the available data that is based on recordings using the XSens motion suit, insoles pressure sensors as well as an eye tracker.

• course A
• course B
• course C

Setup

• Download the data set
• Create a new virtual environment (we name it my_venv) and activate it

python3 -m venv ./my_venv
source ./my_venv/bin/activate

• Install requirements

pip install -r requirements.txt

Generate pandas data frame from .csv files

We provide a script that generates one pandas data frame stored as pickle file from the single recording .csv files. This is quite handy for further processing or analysis.
PYTHONPATH=$(pwd) python src/data_frame_extractor.py -s 1,2 -c a,b -p xxxx/NEWBEE_dataset/data_set/ -d destination_file_path

• -s 1,2 for subjects 1 and 2 (default is all subjects)
• -c a,b for course A and B (default is all courses A,B,C)
• -p path to the data set
• -d path to the data frame stored as pickle file (more parameters are available see python src/data_frame_extractor.py --help)

Labeling tool

The tool allows to inspect the data but also to change the labels or use even custom labels.

It can be started by:
PYTHONPATH=$(pwd) python src/labeling_tool.py -s 1 -c a -p xxxx/NEWBEE_dataset/data_set/

• -s 1 for subject 1
• -c a for course A
• -p path to the data set