This repository contains all the code written and used during a bachelor's thesis about Usability driven design of learning algorithms on a wearable consumer-grade brain-computer interface
. Three different programming languages are used, javascript, C++, and python. Experimental comparisons are conducted, and a javascript based Smartwatch application is implemented. Main programming language is javascript; knowledge in C++ is helpful to understand the full digitial signal processing pipeline; the ecosystem of python is used
to accelerate experiments.
Author: Daniel Bin Schmid
Degree: B.Sc. Medieninformatik
Mail: [email protected]
Chair & supervisor: ITI/ STAR at Universität Stuttgart; Jun.- Prof. Hussam Amrouch
Content is structured in four subfolders.
-
cpp_env
- C++ code. Digital signal processing with iir1, and Riemannian Geometry of covariance matrices with RigByCpp -
node_env
- NodeJs experiments with Node shell. Uses CUDA GPU to accelerate experiments. TensorflowJs is used to facilitate machine learning experiments. -
py_env
- Machine learning experiments with python -
watch_webapp
- The Smartwatch application, written in javascript. Implemented with Vue, Vuetify, and Cordova.
Remember to init submodules by
git submodule update --init --recursive
To download the datasets and move them into their according folders run the bash script download_data.sh