Support Vector Machine Based Approach for Speaker Characterization (2012)
Abstract
This doctoral thesis focuses on the development of algorithms of speaker characterisation by voice. Namely, characterisation of speaker’s identity, and the emotional state detectable in his voice while using the application of state-of-the art classifier algorithm Support Vector Machine (SVM) will be discussed. The first part deals with the state of the art SVM classifier utilised for classification experiments where we search for more sophisticated form of SVM model parameters selection. Also, we successfully apply optimization methods (PSO and GA algorithm) on two classification problems. The second part of this thesis deal with emotion recognition in continuous/dimensional space.
support vector machine – emotion recognition – continuous emotion recognition – SVM – IEMOCAP – automatic emotion recognition
Information
- Author
- Hric, Martin
- Institution
- University of Žilina
- Supervisor
- Publication Year
- 2012
- Upload Date
- May 16, 2013
The current layout is optimized for mobile phones. Page previews, thumbnails, and full abstracts will remain hidden until the browser window grows in width.
The current layout is optimized for tablet devices. Page previews and some thumbnails will remain hidden until the browser window grows in width.