Thesis/Project Openings

Open Bachelor’s Theses

We are developing a learning game based on the Microsoft Kinect camera. Students use the Kinect to browse volumetric data and mark region of interests.

Tasks include:

  • Reimplement the OrganChallenge framework in .NET
  • Implement additional games
  • Evaluate usability and learning effectiveness


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de

The IMCAR at the Uniklinik RWTH Aachen is developing a novel microscopy modality for imaging muscle traction. In this thesis, the muscle traction captured on the microscopy images should be analyzed and quantified in a meaningful way (i.e., flow charts, etc).


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de

In this project, data recorded from a smart shirt will be analyzed on a connected smartphone regarding stress, fatigue and posture. The shirt records basic vital parameters through inexpensive sensors (audio, acceleration)
The tasks include:

  • Processing of the raw input signal to physiological parameters (heart rate, posture, motion)
  • Optimization of the bluetooth transmission protocol
  • Live-evaluation of basic parameters regarding possible overload (posture, stress, fatigue) using existing methods (heart rate variability, …)

For details visit http://mhealth.imib.rwth-aachen.de/hobs/


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de

In this thesis we will explore the possibilities of deep learning to detect abnormalities in human electroencephalogram (EEG). The methods currently used in natural language processing such as recurrent neural networks (RNN) have a high application potential for time series data such as EEG.


Open Master’s Theses

Open Student Projects

In this work, a mobile EEG headset (EMOTIV Epoc+) will be used to differentiate between positive and negative emotions in real-time. Based on existing recording applications, a real-time analysis extension will be developed that processes the EEG data in real-time and detects whether emotions are positive or negative using audio stimuli.


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de

Hospital bed occupancy forecasting is essential for the hospital management. Our group has developed an effective forecasting algorithm, currently implemented in Matlab. The goal of this project is the development of a stand-alone application for use in hospital bed planning. The algorithm involves usage of recurrent neural networks, therefore basic knowledge of machine learning methods is desirable.


As a part of our research on the possibilities of mobile EEG devices in medical context, a mobile visual programming framework for signal processing was developed for Android. This project will focus on adding different signal processing packages, visualization, etc. to the developed framework. Interest in visual programming, brain research and signal processing are welcome.


Psychological disorders are very common; in Europe, more than one in three is affected. As with other psychiatric disorders, early detection of symptoms is also crucial in depression, to minimize personal, social and economic stress. The aim of Psychologist in a Pocket is to use the potential of mobile technology and novel physical sensors in the field of mental health in order to meet the challenge of timely and sensitive detection of the symptoms of depression, in particular when writing. While a lexicon for English and Tagalog already exists, you will collect initial data for a German depression lexicon, for example, using text-mining techniques.


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de

Automatic spike detection is necessary for the analysis of long term electroencephalographic recordings. The goal of the project is to implement some of the know algorithms in MATLAB and evaluate them on partially labelled data.


This is an exploratory project aiming into application of deep learning algorithms to EEG abnormalities classification. The goal is to build a MATLAB framework based on convolutional network, able to outperform the currently existing solution.


Currently, the clinical trials aimed in evaluating the performance of mobile EEG device (Emotiv Epoc) are in progress. EEG signal will be recorded simultaneously by stationary and mobile devices. MATLAB framework for the statistical data analysis is required to compare the data.


Case report forms (CRF) are used in clinical trials to report patient data. While modern electronic CRF (eCRF) systems exist, many trials still rely on pen and paper-based CRFs. Thus, our department supports clinicians with eCRF and printable versions thereof.
The topic of this project is the conversion of scanned CRF sheets (based on a known eCRF) into eCRF data. This includes not only OCR but also detection of checked boxes and consistency checks. A user-friendly software needs to be implemented to let a study nurse upload or scan CRF data and cross-check if correct boxes in an existing eCRF have been filled.


KONTAKT
Stephan Jonas
sjonas@mi.rwth-aachen.de