Large Vocabulary Continuous Speech Recognition 6
HomeFull List of Titles
1: ICSLP'98 Proceedings
Keynote SpeechesText-To-Speech Synthesis 1Spoken Language Models and Dialog 1Prosody and Emotion 1Hidden Markov Model Techniques 1Speaker and Language Recognition 1Multimodal Spoken Language Processing 1Isolated Word RecognitionRobust Speech Processing in Adverse Environments 1Spoken Language Models and Dialog 2Articulatory Modelling 1Talking to Infants, Pets and LoversRobust Speech Processing in Adverse Environments 2Spoken Language Models and Dialog 3Speech Coding 1Articulatory Modelling 2Prosody and Emotion 2Neural Networks, Fuzzy and Evolutionary Methods 1Utterance Verification and Word Spotting 1 / Speaker Adaptation 1Text-To-Speech Synthesis 2Spoken Language Models and Dialog 4Human Speech Perception 1Robust Speech Processing in Adverse Environments 3Speech and Hearing Disorders 1Prosody and Emotion 3Spoken Language Understanding Systems 1Signal Processing and Speech Analysis 1Spoken Language Generation and Translation 1Spoken Language Models and Dialog 5Segmentation, Labelling and Speech Corpora 1Multimodal Spoken Language Processing 2Prosody and Emotion 4Neural Networks, Fuzzy and Evolutionary Methods 2Large Vocabulary Continuous Speech Recognition 1Speaker and Language Recognition 2Signal Processing and Speech Analysis 2Prosody and Emotion 5Robust Speech Processing in Adverse Environments 4Segmentation, Labelling and Speech Corpora 2Speech Technology Applications and Human-Machine Interface 1Large Vocabulary Continuous Speech Recognition 2Text-To-Speech Synthesis 3Language Acquisition 1Acoustic Phonetics 1Speaker Adaptation 2Speech Coding 2Hidden Markov Model Techniques 2Multilingual Perception and Recognition 1Large Vocabulary Continuous Speech Recognition 3Articulatory Modelling 3Language Acquisition 2Speaker and Language Recognition 3Text-To-Speech Synthesis 4Spoken Language Understanding Systems 4Human Speech Perception 2Large Vocabulary Continuous Speech Recognition 4Spoken Language Understanding Systems 2Signal Processing and Speech Analysis 3Human Speech Perception 3Speaker Adaptation 3Spoken Language Understanding Systems 3Multimodal Spoken Language Processing 3Acoustic Phonetics 2Large Vocabulary Continuous Speech Recognition 5Speech Coding 3Language Acquisition 3 / Multilingual Perception and Recognition 2Segmentation, Labelling and Speech Corpora 3Text-To-Speech Synthesis 5Spoken Language Generation and Translation 2Human Speech Perception 4Robust Speech Processing in Adverse Environments 5Text-To-Speech Synthesis 6Speech Technology Applications and Human-Machine Interface 2Prosody and Emotion 6Hidden Markov Model Techniques 3Speech and Hearing Disorders 2 / Speech Processing for the Speech and Hearing Impaired 1Human Speech ProductionSegmentation, Labelling and Speech Corpora 4Speaker and Language Recognition 4Speech Technology Applications and Human-Machine Interface 3Utterance Verification and Word Spotting 2Large Vocabulary Continuous Speech Recognition 6Neural Networks, Fuzzy and Evolutionary Methods 3Speech Processing for the Speech-Impaired and Hearing-Impaired 2Prosody and Emotion 72: SST Student Day
SST Student Day - Poster Session 1SST Student Day - Poster Session 2Author Index
A B C D E F G H I
J K L M N O P Q R
S T U V W X Y Z
Multimedia Files
Sharable Software Repository for Japanese Large Vocabulary Continuous Speech Recognition
Authors:
Tatsuya Kawahara, Kyoto Univ. (Japan)
Tetsunori Kobayashi, Waseda Univ. (Japan)
Kazuya Takeda, Nagoya Univ. (Japan)
Nobuaki Minematsu, Toyohashi Univ. of Tech. (Japan)
Katsunobu Itou, ETL (Japan)
Mikio Yamamoto, Tsukuba Univ. (Japan)
Atsushi Yamada, ASTEM (Japan)
Takehito Utsuro, Nara Institute of Science and Technology (Japan)
Kiyohiro Shikano, Nara Institute of Science and Technology (Japan)
Page (NA) Paper number 763
Abstract:
The project of Japanese LVCSR (Large Vocabulary Continuous Speech Recognition) platform is introduced. It is a collaboration of researchers of different academic institutes and intended to develop a sharable software repository of not only databases but also models and programs. The platform consists of a standard recognition engine, Japanese phone models and Japanese statistical language models. A set of Japanese phone HMMs are trained with ASJ (Acoustic Society of Japan) databases of 20K sentence utterances per each gender. Japanese word N-gram (2-gram and 3-gram) models are constructed with a corpus of Mainichi newspaper of four years. The recognition engine JULIUS is developed for assessment of both acoustic and language models. The modules are integrated as a Japanese LVCSR system and evaluated on 5000-word dictation task. The software repository is available to the public.
SL980763.PDF (From Author)
SL980763.PDF (Rasterized)
TOP
The Design of the Newspaper-Based Japanese Large Vocabulary Continuous Speech Recognition Corpus
Authors:
Katunobu Itou, ETL (Japan)
Mikio Yamamoto, Univ. of Tsukuba (Japan)
Kazuya Takeda, Nagoya Univ. (Japan)
Toshiyuki Takezawa, ATR (Japan)
Tatsuo Matsuoka, NTT (Japan)
Tetsunori Kobayashi, Waseda Univ. (Japan)
Kiyohiro Shikano, NAIST (Japan)
Shuichi Itahashi, University of Tsukuba (Japan)
Page (NA) Paper number 722
Abstract:
In this paper we present the first public Japanese speech corpus for large vocabulary continuous speech recognition (LVCSR) technology, which we have titled JNAS (Japanese Newspaper Article Sentences). We designed it to be comparable to the corpora used in the American and European LVCSR projects. The corpus contains speech recordings (60 hrs.) and their orthographic transcriptions for 306 speakers (153 males and 153 females) reading excerpts from the newspaper's articles and phonetically balanced (PB) sentences. This corpus contains utterances of about 45,000 sentences as a whole with each speaker reading about 150 sentences. JNAS is being distributed on 16 CD-ROMs.
SL980722.PDF (From Author)
SL980722.PDF (Rasterized)
TOP
Indexing and Classification of TV News Articles Based on Speech Dictation Using Word Bigram
Authors:
Jun Ogata, Ryukoku University (Japan)
Yasuo Ariki, Ryukoku University (Japan)
Page (NA) Paper number 126
Abstract:
In order to construct a news database with a function of video on demand (VOD), it is required to classify news articles into topics. In this paper, we propose a method to automatically index and classify TV news articles into 10 topics based on a speech dictation techniques using speaker independent triphone HMMs and word bigram.
SL980126.PDF (From Author)
SL980126.PDF (Rasterized)
TOP
Parametric Trajectory Mixtures for LVCSR
Authors:
Man-Hung Siu, GTE/BBN Technologies (USA)
Rukmini Iyer, GTE/BBN Technologies (USA)
Herbert Gish, GTE/BBN Technologies (USA)
Carl Quillen, GTE/BBN Technologies (USA)
Page (NA) Paper number 890
Abstract:
Parametric trajectory models explicitly represent the temporal evolution of the speech features as a Gaussian process with time-varying parameters. HMMs are a special case of such models, one in which the trajectory constraints in the speech segment are ignored by the assumption of conditional independence across frames within the segment. In this paper, we investigate in detail some extensions to our trajectory modeling approach aimed at improving LVCSR performance: (i) improved modeling of mixtures of trajectories via better initialization, (ii) modeling of context dependence, and (iii) improved segment boundaries by means of search. We will present results in terms of both phone classification and recognition accuracy on the Switchboard corpus.