Difference between revisions of "Online Emotion Classification"

From AIRWiki
Jump to: navigation, search
m
m
 
(One intermediate revision by one other user not shown)
Line 1: Line 1:
 
{{Project
 
{{Project
 
|title=Emotion from Interaction
 
|title=Emotion from Interaction
|coordinator=SimoneTognetti
+
|coordinator=AndreaBonarini
 +
|tutor=SimoneTognetti
 
|students=AndreaMaesani;ClaudioMagni;EmanuelePadula
 
|students=AndreaMaesani;ClaudioMagni;EmanuelePadula
 
|resarea=Affective Computing
 
|resarea=Affective Computing
 
|restopic=Affective Computing And BioSignals
 
|restopic=Affective Computing And BioSignals
 
|start=2008/10/07
 
|start=2008/10/07
|end=2009/06/01
+
|end=2009/06/02
 
|status=Closed
 
|status=Closed
 
|level=Ms
 
|level=Ms

Latest revision as of 16:58, 3 October 2011

Emotion from Interaction
Coordinator: AndreaBonarini (andrea.bonarini@polimi.it)
Tutor: SimoneTognetti (tognetti@elet.polimi.it)
Collaborator:
Students: AndreaMaesani (andrea.maesani@mail.polimi.it), ClaudioMagni (claudio.magni@mail.polimi.it), EmanuelePadula (e.padula@gmail.com)
Research Area: Affective Computing
Research Topic: Affective Computing And BioSignals
Start: 2008/10/07
End: 2009/06/02
Status: Closed
Level: Ms
Type: Course

Project description

The project focuses on the development of a software framework for online emotion classification. A general framework will be developed to support emotion detection using several biometric signals.

The framework for online emotion classification will be composed of:

  • A client application that receives data (XML data) from a generic source (A server application that polls sensors and generate the XML data with the sampled biometric signals). [C++/libXML/MatlabAPI]
  • An online classifier that can be feed with the data received by the client and outputs the result of the classification. [Matlab/Java]
  • An offline trainer to train the classifier with previously collected data. [Matlab/Java]
  • A 3d GUI that can be connected to the online classifier to show the result of the classification. [C++/OGRE]

The client parses the received XML and interacts through the Matlab API with the Matlab engine, launching the classifier and passing it the parsed data.

The classifiers are based on the Weka classification engine. After a strong preprocessing on the data done with Matlab, the signals are classified using Weka.

The GUI contains several models of human faces that can change expressions according to the received classification signal from the online classifier.