This class explains the layout of state-of-the-art speech recognition systems. The layout will be motivated based on the human speech production process und its properties. The class treats all processing steps of automatic speech recognition systems in detail: signal pre-processing, training of suitable, statistical models, and the actual recognition process. The focus will be on statistical methods, as they are being used in current speech recognition systems. In this way the state-of-the-art of the area of automatic speech recognition will be communicated. Further the class will introduce alternative Methods, which were the foundation of the current methods and which are still being used in special circumstances. Using sample applications und examples from current research projects, the current state-of-the-art and the performance of current systems will be illustrated.

• Xuedong Huang, Alex Acero, Hsiao-wuen Hon, Spoken Language Processing, Prentice Hall, NJ, USA, 2001
• Fredrick Jelinek (editor), Statistical Methods for Speech Recognition, The MIT Press,1997, Cambridge, Massachusetts, London, England

Weiterführende Literatur

• Lawrence Rabiner and Ronald W. Schafer, Digital Processing of Speech Signals, Prentice Hall, 1978
• Schukat-Talamazzini, Automatische Spracherkennung

Die Vorlesung wird über Zoom gehalten werden.

Nähere Informationen dazu gibt es im ILIAS Kurs.

Zoom-Link:https://zoom.us/j/92732046306