The SpeechLab group is a part of the Institute of Information Technology and Electronics. It was founded in 1994 by prof. Jan Nouza who has been working in the field of speech processing since his university studies in 1980s. Recently, the team consists of 10 researchers, all having their Ph.D. degrees either in speech or image processing.
The picture mosaic below depicts some of the previous interesting projects we were working on, such as: the first ASR system for Czech language (from 1990s and early 2000s), system VISPER for visualizing and experimenting with HMMs (Hidden Markov Models), several examples of voice dialogue systems, a talking face, a program for voice control of a PC developed for users with specific needs, etc.
Since 1996, our team participated in more than 15 large research projects supported by Czech research grant agencies (GACR and TACR), or by Czech ministries (e.g. Ministry of Education, Ministry of Interior and Ministry of Culture. We were active partners in several European COST (Cooperation in Science and Technology) projects, too.
Here is a list of the most interesting research projects solved in the recent decade (2011-2020):
- Speech-To-Text Technology to Transcribe and Disclose 100,000+ Hours of Bilingual Documents from Historical Czech and Czechoslovak Radio Archive (4-year project supported by Czech Ministry of Culture) – see our Interspeech paper [1]
- Development of On-Line Multilingual Broadcast Transcription System for 13 Slavic Languages (supported by 2 large research projects of Technology Agency of the Czech Republic), see publications [2, 3, 4]
- Development of a Very Fast Keyword Spotting System for On-line and Off-line Speech Mining Applications (supported by a project of Technology Agency of the Czech Republic), see publication [5]
- Neural Network based methods and modules for Event Identification in Speech Signal (Voice Activity Detectors, Speaker Changer Detectors, Language Identification Detectors, etc.), see publications [6, 7, 8]
And a list of some commercial products based on our research and development. Most of them are used and sold by Newton Technologies company:
- Newton Dictate program – professional voice dictation software, currently available for Czech, Slovak, Polish, Croatian and Serbian
- Multilingual On-Line Broadcast Transcription System – runs in 24/7 mode and monitors more than 50 TV and radio stations in 20 European languages
- Speech Processing and Analysis Tools for Call Centres – provided by Newton Technologies for several call centres in Czechia and abroad
- MyVoice or MyDictate. – voice controlled PC applications for users with specific needs
Are you a student, a postdoc or a commercial subject interested in our research? We will appreciate if you contact us:
jan.nouza AT tul.cz, petr.cerva AT tul.cz