AuBT - Augsburg Biosignal Toolbox
|Projektverantwortung vor Ort:||M.Sc. Johannes Wagner|
Augsburg Biosignal Toolbox (AuBT) provides tools to analyse physiological signals in the face of emotion recognition. It consists of a bunch of Matlab functions which have been developed to meet the following tasks:
• extract features from physiological signals
• automatically select the relevant features
• use these features to train and evaluate a classifier
A graphical user interface (AuBTgui) serves as interface between user and toolbox. It provides an easy and fast way to analyze new data and gives access to all functions without a long period of vocational adjustment.
AuBT comes along with two corpora: a corpus containing physiological data of a single user in four different emotional states and a corpus containing physiological data recorded from a single user under varying stress.
Important: to run the graphical interface Matlab 7 or higher is necessary. However, we are trying to build a version that is also compatible with older Matlab versions.
You can download the software from:
and the documentation from:
A tutorial video is available from:
- 2006/11/20 changed "aubt_normRange.m": supports now a matrix as input (with old version an error occured when using MLP classifier)
- 2006/12/13 changed "aubt_classifier.m": data is now normalized before KNN is used for classification
- 2006/12/15 changed "AuBTGui.m": a problem has been fixed that occured when swapping features in the boxplot window
- 2007/02/22 changed "ShowFeat.m": another problem has been fixed that occured during swapping of features in the boxplot window
- 2007/10/29 changed "aubt_lowpassFilter.m": offers now option to do causal filtering
- 2009/07/30 changed "aubt_extractFeatECG.m": made compatible with Matlab versions newer than 7.1
- 2013/02/15 changed "aubt_extractFeatECG.m": bug fix
- 2013/02/15 changed "aubt_extractFeatSC": bug fix (thanks to David Gasul)
- Johannes Wagner, Jonghwa Kim, Elisabeth André. From Physiological Signals to Emotions: Implementing and Comparing Selected Methods for Feature Extraction and Classification. In IEEE International Conference on Multimedia & Expo (ICME 2005), 2005.
- Jonghwa Kim, Elisabeth André, Matthias Rehm, Thurid Vogt, Johannes Wagner. Integrating Information from Speech and Physiological Signals to Achieve Emotional Sensitivity. In Proc. of the 9th European Conference on Speech Communication and Technology, 2005.
- Johannes Wagner. Vom physiologischen Signal zur Emotion: Implementierung und Vergleich ausgewählter Methoden zur Merkmalsextraktion und Klassifikation, Technical Report, Institute of Computer Science, University of Augsburg, 2005. [GERMAN]