Principal component analysis for emotion recognition

OData support
Supervisor:
Dr. Vicsi Klára
Department of Telecommunications and Media Informatics

In the past semesters during „Individual Laboratory” I participated in the improvement of the the existing emotion – database, I tried to get familiar with the speech recognizer machine (SVM) and the program called Praat. I also dealt with the examination of recognition of emotions on sentence-based speech units.

With this knowledge I will attempt a new procedure, the principal component analysis which is used to determine which acoustic parameters are relevant to emotion recognition. This will hopefully help further optimization opportunities and we can achieve better results. To fulfill my task I used the database based on emotions provided by the Speech Technology Laboratory.

In the first part of my thesis I discuss generally about emotions, then I turn to emotion database.

The biggest part of my work deals with description of the basis of principal component analysis and it’s practical implementation. The implementation is done in Matlab, I illustrate the steps.

At the end of the thesis I interpret the results and write about my gathered experience.

Downloads

Please sign in to download the files of this thesis.