Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/184856
Title: Robust Feature Extraction Techniques for Speech Recognition using Autocorrelation Domain Processing
Researcher: Poonam Bansal
Guide(s): Shail Bala Jain and Amita Dev
Keywords: 
University: Guru Gobind Singh Indraprastha University
Completed Date: 2010
Abstract: Automatic speech recognition (ASR) is one of the most important research areas in the field of speech technology and research. It is also known as the recognition of speech by a machine or, by some artificial intelligence. However, in spite of focused research in this field for the past several decades, robust speech recognition with high reliability has not been achieved as it degrades in the presence of speaker variations, channel mismatch conditions, and in noisy environments. Significant advances have been made in these directions but the performance of current solutions in the presence of ambient acoustic noise is still a research problem in the speech recognition area. One of the major factors for the overall deficiency of any ASR system is due to the use of poor feature extraction methods. This thesis aims to develop techniques to extract robust features from speech so as to achieve improved performance in speech recognition. newlineContributions made in this thesis are aimed at improving the performance of ASR system in the presence of ambient acoustic noise by robust feature extraction. As the production mechanisms of speech and noise signals are different, transforming them into autocorrelation domain provides a better representation for noise robust processing. In this thesis, feature extraction techniques are developed for improving the noise robustness of feature extraction algorithms based on processing the degraded speech signal in the autocorrelation domain. newlineThe autocorrelation domain is well-known for its pole preserving and noise separation properties. Firstly, a robust feature extraction technique using differentiated relative autocorrelation sequence spectrum (DRASS) is developed.
Pagination: 
URI: http://hdl.handle.net/10603/184856
Appears in Departments:University School of Engineering and Technology



Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.