Noise Robust Automatic Speech Recognition

ISBN-13: 9783639149609 / Angielski / Miękka / 2009 / 172 str.

Hemant Misra

Opis

The recent advancements in automatic speechrecognition (ASR) systems have expedited their commercial deployment.Still, a major challenge faced by the current ASR systems is their poor performance for any mismatch between training andtesting conditions. This book presents an in-depth analysisof multi-stream combination approach to improve the robustnessof ASR systems for mismatch caused by additive noise. Thebook also discusses that multi-stream approach is complementaryto other approaches for noise robustness, and how they can beused together. The selection of feature streams andallocation of weights to each feature stream are two important issues inmulti-stream. Both the issues are analyzed, and a step-wiseapproach is presented to realize a significant improvement inperformance by a multi-stream system compared to that of astate-of-the-art system. The analysis proposed in this work can also be applied to other pattern recognitiontasks such as speaker recogntion and text-processing, renderingthis book useful for a wide audience ranging from beginner to advancestage users working in ASR as well as other pattern recognitiondomains.

O autorze

Dr. Hemant Misra is an active researcher in the areas of
automatic speech/speaker
recognition, text processing and machine learning, and is a
post-doc researcher at
University of Glasgow. His academic and industry research career
spans more than a
decade. He completed M.S. and Ph.D. from IIT Madras, India, and
EPFL, Switzerland,
respectively.