Normalization And Transformation Techniques For Robust Speaker Recognition
Download Normalization And Transformation Techniques For Robust Speaker Recognition PDF/ePub or read online books in Mobi eBooks. Click Download or Read Online button to get Normalization And Transformation Techniques For Robust Speaker Recognition book now. This website allows unlimited access to, at the time of writing, more than 1.5 million titles, including hundreds of thousands of titles in various foreign languages.
Normalization and Transformation Techniques for Robust Speaker Recognition
In this section, we shall comment, mainly from two different perspectives, capacities and limitations of normalization and transformation methods at both score and feature levels. In one way, we discuss how these techniques help to reduce mismatch between training and test conditions is considered. In the second one, we discuss how these methods are combined with other parts of SR systems, as either early or late processing module. 5.1 Mismatch reduction How to deal with noises is a universal topic for robust speaker recognition. All the methods reviewed in this chapter are absolutely concerned with this topic. However, noises can normally be categorized into two groups, i.e., additive noise and convolutional noise. The additive noise further consists of environmental noise and cross-talking from other speakers, while convolutional noise is mainly caused by communication channels. For most noise, it occurs in a sudden and unpredictable way. Its interference results in a huge mismatch from the models trained before-hand, which in turn often severely degrades system performance. Thus, for robust speaker recognition, one of the key problems is how to reduce mismatch between different training and test scenarios. Of course, mismatch always exits, no matter how hard a system is carefully prepared. So a smart system knows how to compensate a variety of mismatches, whereas a poor system does not. We shall see, in such a context, how well each of the normalization methods works in terms of mismatch reduction.
Speech Recognition
Author: France Mihelič
language: en
Publisher: BoD – Books on Demand
Release Date: 2008-11-01
Chapters in the first part of the book cover all the essential speech processing techniques for building robust, automatic speech recognition systems: the representation for speech signals and the methods for speech-features extraction, acoustic and language modeling, efficient algorithms for searching the hypothesis space, and multimodal approaches to speech recognition. The last part of the book is devoted to other speech processing applications that can use the information from automatic speech recognition for speaker identification and tracking, for prosody modeling in emotion-detection systems and in other speech processing applications that are able to operate in real-world environments, like mobile communication services and smart homes.
Robust Speaker Recognition in Noisy Environments
This book discusses speaker recognition methods to deal with realistic variable noisy environments. The text covers authentication systems for; robust noisy background environments, functions in real time and incorporated in mobile devices. The book focuses on different approaches to enhance the accuracy of speaker recognition in presence of varying background environments. The authors examine: (a) Feature compensation using multiple background models, (b) Feature mapping using data-driven stochastic models, (c) Design of super vector- based GMM-SVM framework for robust speaker recognition, (d) Total variability modeling (i-vectors) in a discriminative framework and (e) Boosting method to fuse evidences from multiple SVM models.