Head: Dipl. Ing. Milan Rusko
Object of the research :
a) theoretical part
Trainable (self-learning) automated speech processing systems
b) application part
- Speech recognizers building for the applications in telecommunications, robotics, medicine, banking etc. Design of the modern recognition systems requires building of the training speech-databases. The recognizers can form a part of the automated translation systems (machine translation) and the input interface for the voice-driven teleservices.
- Developing speech synthesis systems with higher naturalness, reaching compatibility with the mostly used operation systems (Windows, Linux, Unix) and also with Java. Development of systems for disabled or blind people. Building of special training databases for modern data-driven high-level speech synthesis systems.
c) additional areas of research
speech acoustics and musical acoustics