Диплом (Master), Massachusetts Institute of Technology, 2008. — 103 p. Discriminative training for acoustic models has been widely studied to improve the performance of automatic speech recognition systems. To enhance the generalization ability of discriminatively trained models, a large-margin training framework has recently been proposed. This work investigates large-margin...
Диплом (Master), Massachusetts Institute of Technology, 1996, -70 pp. The objective of this research is to investigate the use of a hierarchical framework for phonetic classification of speech. The framework is motivated by the observation that a wide variety of measurements may be needed to make phonetic distinctions among different types of speech sounds. The measurements...
Диссертация, Katholieke Universiteit Leuven, 1998. — 218 p. In general the aim of an automatic speech recognition system is to write down what is said. State of the art continuous speech recognition systems for large vocabulary - for an entire language - consist of four basic modules: the signal processing, the acoustic modelling, the language modelling and the search engine....
Диссертация (Master), Massachusetts Institute of Technology, 1992, -97 pp.
The Viterbi search is an important but computationally expensive algorithm for speech recognition. Even with the substantial advances expected in processor technology the massive computational resources required will remain prohibitive for operation of a speech recognition system in real time. This...
Диплом (Magister), Universität Wien, 2009, -126 pp. In recent years the number of court cases involving speech recordings of suspects as evidence, for example taken from telephone conversations, has seen a substantial increase. Forensic speech evidence is expected to gain even more importance, as speech communication technologies have become ubiquitous. Likewise the role of...
Диссертация (Master), University of Cambridge, 1997, -67 pp. This project investigates the problem of labelling segments in a speaker-tracking system. A mathematical representation of each segment is sought which encaptures the speaker-dependent information available. It is shown that both the covariance matrix and the Maximum Likelihood Linear Regression (MLLR) matrix provide...
Master Thesis. — Eidgenössische Technische Hochschule Zürich, 2003. — 119 p. Large vocabulary speech recognition systems traditionally represent words in terms of smaller subword units. During training and recognition they require a mapping table, called the dictionary, which maps words into sequences of these subword units. The performance of the speech recognition system...
Диплом (Master), Massachusetts Institute of Technology, 1996, -65 pp. In an effort to reduce the degradation in speech recognition performance caused by variations in vocal tract shape among speakers, this thesis studies a set of low complexity, maximum likelihood based speaker normalization procedures. By approximately modeling the vocal tract as a simple acoustic tube, these...
Диплом (Master), Helsinki University of Technology, 1999, -113 pp. Synthetic or artificial speech has been developed steadily during the last decades. Especially, the intelligibility has reached an adequate level for most applications, especially for communication impaired people. The intelligibility of synthetic speech may also be increased considerably with visual...
Диссертация (Master), China University of Science and Technology, 1996. — 78 p. This research identifies the problems encountered in transmitting voice over the Internet and proposes approaches to solve these problems. The current Internet is not very suitable for transmitting real-time data because its underlying protocols and switches were only engineered to transmit non-real...
Диплом (Master), Mississippi State University, 2008. — 70 p. In this work, nonlinear acoustic information is combined with traditional linear acoustic information to produce a noise-robust feature set for speech recognition. Classical acoustic modeling has relied on the assumption of linear acoustics where signal processing is performed in the signal's frequency domain....
Диплом (Master), Universität Karlsruhe, 2006, -86 pp. The following report describes the work at the interAct on Bulgarian speech recognition, including the collection of data, training a Bulgarian speech recognizer and experimenting with Russian text data to improve the recognition. It also gives an overview of the unique traits of Bulgarian language, introduces the main...
Диплом (Master), Indian Institute of Technology, 2001. — 85 p. The thesis presents a novel situationally-aware multimodal spoken language system called Fuse that performs speech understanding for visual object selection. Fuse uses semantic information from immediate visual context to guide spoken language recognition and understanding. An experimental task was created in which...
Master report, Carnegie Mellon University, 2005, -46 pp. Given a speech signal there are two kinds of information that may be extracted from it. On one hand there is the linguistic information about what is being said, and on the other there is also speaker specific information. This report deals with the task of speaker recognition where the goal is to determine which one of a...
Диплом (Master), Mississippi State University, 2000, -135 pp. Over the past few years, speech recognition technology performance on tasks ranging from isolated digit recognition to conversational speech has dramatically improved. Performance on limited recognition tasks in noise-free environments is comparable to that achieved by human transcribers. This advancement in...
Диплом (Master), University of Illinois, 2010. — 91 p. In this thesis, we describe a biometric authentication system that is capable of recognizing its users’ voice using advanced machine learning and digital signal processing tools. The proposed system can both validate a person’s identity (i.e. verification) and recognize it from a larger known group of people (i.e....
Диссертация (Master), University of Cambridge, 1999, -42 pp.
Most if not all speech recognition systems use Hidden Markov Models (HMM) to model the production of speech from sequences of phones or other basic units of speech. HMMs need to be trained, and this is done using speech utterances whose transcrip t ion is known. The most common method of t raining HMMs is known as...
Диссертация (Master), Wilfrid Laurier University, 1989, -177 pp. Speech Recognition is a rapidly expanding field with many useful applications in man-machine interfacing. One of the main benefits of speech control is the flexibility and ease of use allowed an operator for any number of specific applications. Speech recognition units (SRU) are currently at a high level of...
Диссертация (Master), Universitetet i Trondheim, 1994, -97 pp. Automatic recognition of speech has come a long way from the first serious attempts at machine recognition of a few isolated words in the 1950's. Today, commercial recognizers capable of recognizing several tens of thousands words spoken as isolated utterances are available on a PC platform and the first speaker...
Диплом (Master), Helsinki University of Technology, 2007, -66 pp. The duration of phones play a significant part in the comprehension of speech. Finnish, for example, has several word pairs which can be distinguishable mainly by the duration of their phones. In automatic speech recognition, it is very important to detect these differences. Modern speech recognition systems,...
Диплом (Master), Mississippi State University, 2002, -91 pp. Rapid advances in speech recognition theory, as well as computing hardware, have led to the development of machines that can take human speech as input, decode the information content of the speech, and respond accordingly. Real-time performance of such systems is often dominated by the evaluation of likelihoods in...
Диплом (Master), Mississippi State University, 2006. — 94 p. Early human language technology systems were designed in a monolithic fashion. As these systems became more complex, this design became untenable. In its place, the concept of distributed processing evolved wherein the monolithic structure was decomposed into a number of functional components that could interact...
Диплом (Master), Mississippi State University, 2003, -62 pp. Supervised learning using Hidden Markov Models has been used to train acoustic models for automatic speech recognition for several years. Typically clean transcriptions form the basis for this training regimen. However, results have shown that using sources of readily available transcriptions, which can be erroneous...
Диплом (Master), Massachusetts Institute of Technology, 1998, -77 pp. In this thesis, eigenstructure based noise suppression techniques are developed to improve the performance of LPC spectral estimation of speech signals in the presence of additive white noise. LPC estimation error increases as the SNR of the speech signal decreases, thus affecting the performance of speech...
Диссертация (Master), McGill University, 2000. — 53 p. Automatic speech recognition by machine has been a goal of speech researchers for more than 40 years. In recent years we have seen great advances in speech recognition technology. Some speech recognition techniques have entered into the market place and been used in applications such as command-and-control, credit-card...
Работа на получение квалификационного уровня магистра электроники.
Специальность: 8.090803 - Электронные системы.
Научный руководитель - Велигорский Александр Анатольевич.
Чернигов, 2004.
Разделы:
Аналитический обзор методов речевого кодирования.
Принципы построения систем с адаптивной дифференциальной импульсно-кодовой модуляцией.
Влияние параметров системы на качество...
Дипломна робота (Магістр), Національній авіаційний університет. — Київ, 2014. — 107 с. Спеціальність 8.05090302 «Телекомунікаційні системи та мережі» Науковий керівник: д-р техн. наук, проф. Давлет'янц О.І. В дипломній роботі вирішуються актуальні проблеми стиснення мовних сигналів та шляхи підвищення її якості. Були розроблені алгоритми підвищення якості стиснення мовних...
Санкт-Петербургский политехнический университет Петра Великого, Институт компьютерных наук и технологий, Кафедра компьютерных интеллектуальных технологий, Тимонин В.М., Санкт-Петербург, 2015, 80 с. Тема относится к области цифровой обработки речевого сигнала. Предложен и реализован новый алгоритм морфинга человеческого голоса с дополнительными возможностями. Написано приложение,...
Комментарии