Bernd T. Meyer
Hi, welcome to my academic home!
Medical Physics Section
Phone: +49 441 798-3528
Fax: +49 441 798-3902
Office location: W2 0-084b
Note: substitute _AT_ by @ within the email address, please
- The combination of machine learning and speech
- Auditory feature extraction for automatic speech recognition (ASR)
- Computational Auditory Scene Analysis (CASA)
- Using machine learning for models of human speech perception
- Comparison of human and automatic speech recognition
- Application of novel feature concepts on tasks in speech research (e.g., speech activity detection or speaker recognition)
I've been involved in the development of the Oldenburg Logatome Corpus (OLLO), which is freely available for research. Follow this link for a description: http://medi.uni-oldenburg.de/ollo. I'm also interested in physiologically motivated spectro-temporal features as a frontend for ASR. Part of this work is described in (Schädler et al., JASA, 2012), and the corresponding piece of code for feature extraction can be downloaded here: http://medi.uni-oldenburg.de/GBFB.