This project contains code written in relation to our bachelor project at NTNU. Our assignment was to classify sitting postures with machine learning algorithms on accelerometer- and gyroscope-data. The datasets are recorded using sensors from LP-RESEARCH, and provided in this repository. Additionally, we have developed an electron app for recording data and classifying your sitting posture in realtime.
- Download latest release
- Open a terminal in the project root
- Run
pip install -r "requirements.txt"
- Run the command
start.bat
for windows orstart.command
for mac
As we depend on another repo for the OpenZen-library, you have to clone the project with the submodules included. The following line will fix this for you:
git clone --recurse-submodules https://github.com/OleJonas/Got-Your-Back.git 'GotYourBack'
You will also need to run npm install
inside the frontend-folder, and pip install -r "requirements.txt"
inside root. For running the application you will need to make it using npm run make_windows
for windows or npm run make
for mac. Another option is to run as development server either in browser (npm start
) or as an electron application (npm run dev
).
If you are experiencing any problems connecting to the sensors (eg. The file "btScan" doesn't exist
), try building the openzen lib again.
This is done by running the following commands:
cd ./lib/openzen
mkdir build && cd build
make -j4
cmake -DZEN_PYTHON=ON ..
In the openzen-folder, you wil now find a file called openzen.pyd (Windows), or openzen.so (Unix). Copy this file and place it in the scripts- and backend-folder.