首页--工业技术论文--无线电电子学、电信技术论文--通信论文--电声技术和语音信号处理论文--语音信号处理论文--语音识别与设备论文

Hidden Markov Model Based Automatic Speech Recognition Using Mel Frequency Cepstral Coefficients in Nepalese

List of Tables第7-8页
List of Figures第8-9页
Abstract第9页
Acknowledgements第10-11页
Chapter 1 Introduction to ASR第11-19页
    1.1 Introduction第11页
    1.2 Are Speech Recognition and Voice Recognition Synonymous?第11-12页
    1.3 Why Speech Recognition? Uses and Applications第12-13页
    1.4 Types of Speech Recognition第13页
    1.5 An overview of ASR第13-14页
    1.6 Project Outline第14-15页
    1.7 Understanding Digital Audio第15-17页
    1.8 The Hidden Markov Model Toolkit (HTK)第17-19页
Chapter 2 Understanding the features of Nepali Language第19-28页
    2.1 Devnagari Script第19-20页
    2.2 Romanization of Nepali words第20-26页
    2.3 Representation of Pronunciation of Nepali Words第26-28页
Chapter 3 Classification of Nepalese phonemes第28-37页
    3.1 What is a phoneme?第28-29页
    3.2 Why classification?第29页
    3.3 Consonants and Vowels第29-37页
        3.3.1 Consonants第30-35页
        3.3.2 Vowels第35-37页
Chapter 4 Hidden Markov Model第37-52页
    4.1 Definition第37-39页
    4.2 Assumptions in the theory of HMMs第39页
    4.3 Three main questions on HMMs第39-52页
        4.3.1 The Evaluation Problem and the Forward Algorithm第40-42页
        4.3.2 The Decoding Problem and the Viterbi Algorithm第42-43页
        4.3.3 The Learning Problem第43-52页
Chapter 5 Understanding Hidden Markov Model Tool Kit (HTK)第52-57页
    5.1 HTK Software Architecture第52-53页
    5.2 The Toolkit第53-57页
        5.2.1 Data Preparation Tools第54-55页
        5.2.2 Training Tools第55页
        5.2.3 Recognition Tools第55-56页
        5.2.4 Analysis Tool第56-57页
Chapter 6 Implementation Details第57-78页
    6.1 Preparing Target Nepali Vocabulary第57-61页
        6.1.1 Initialization第59-60页
        6.1.2 Transcription of Devnagari files第60页
        6.1.3 Making new entries in the dictionary database第60-61页
    6.2 Language Model第61-63页
    6.3 Recording the speech data第63页
    6.4 Creating the Phoneme-level Transcriptions第63-66页
    6.5 Parameterization of speech data第66-67页
    6.6 Creating monophone HMMs第67-70页
    6.7 Creating tied state tri-phone HMMs第70-74页
    6.8 Evaluation第74-76页
    6.9 Problems Encountered第76-78页
References第78-79页

论文共79页,点击 下载论文
上一篇:子宫内膜癌中乙酰肝素酶、组织蛋白酶D和碱性成纤维细胞生长因子的表达
下一篇:Candidate Base Stations a Security Solution for Compromised Base Stations in Wireless Sensor Networks