On Friday, 27 February, Dr Tomohiro Nakatani from the Communication Science Laboratories of NTT Corporation in Japan will be giving a lecture in English at Paderborn University. Under the title "Enhancing Distant Automatic Speech Recognition via Model-Based Multi-Microphone Front-Ends", he will focus on automatic speech recognition with audio signals recorded in the far field. This technology is a key requirement for applications such as recognising natural conversations in real everyday environments. Interference factors such as background noise, reverberation and overlapping speakers pose a particular challenge. The event starts at 11 a.m. in lecture theatre L3.204. Registration is not required. Further information can be found on the website of the Faculty of Computer Science, Electrical Engineering and Mathematics.
Dr Nakatani will give an overview of so-called model-based multi-microphone front-ends, which are used to suppress interfering factors. A key advantage of these methods is their ability to decompose audio signals using physical and probabilistic models - often without prior training. This makes them particularly suitable for unknown and complex acoustic environments. In combination with neural networks, these approaches enable high-precision signal processing even under difficult conditions. Using challenging scenarios of "Distant Automatic Speech Recognition", Dr. Nakatani will show how noise reduction, source separation and reverberation suppression methods can significantly improve recognition performance.
Dr Nakatani is a Distinguished Industry Speaker of the IEEE Signal Processing Society. The host at Paderborn University is Prof. Dr Reinhold Häb-Umbach.
This text was translated automatically.
