Julius LVCSR 2015


Notes:

Large-vocabulary continuous speech recognition (LVCSR, more commonly known as speech-to-text, full transcription or ASR – automatic speech recognition) uses a set of words (bi-grams, tri-grams etc.) as the basic unit.

Resources:

Wikipedia:

See also:

Julius LVCSR 2013Julius LVCSR 2014


Evaluation of advanced language modeling techniques for the Slovak LVCSR D Zlacký, J Staš, J Juhár… – … ), 2015 25th International …, 2015 – ieeexplore.ieee.org … 1549-1552. [15] A. Lee and T. Kawahara, “Recent Development of Open-Source Speech Recognition Engine Julius”, in Proc. … 261-276. [18] J. Staš, D. Hládek, and J. Juhár, “Language Model Adaptation for Slo- vak LVCSR”, in Proc. … Related articles All 3 versions

Morpheme Segmentation and Concatenation Approaches for Uyghur LVCSR M Ablimit, T Kawahara, A Hamdulla – International Journal of Hybrid …, 2015 – sersc.org … ICSP, Beijing, (2010). [2] A. Lee, T. Kawahara, and K. Shikano, “Julius — an open source real-time large vocabulary recognition engine,” In Proc. Eurospeech, pp. … [4] OW. Kwon, “Performance of LVCSR with morpheme-based and syllable-based recognition units,” In Proc. … Related articles All 2 versions

Spoken term detection ALBAYZIN 2014 evaluation: overview, systems, results, and discussion J Tejedor, DT Toledano… – EURASIP …, 2015 – asmp.eurasipjournals.springeropen. … … Julian David Echeverry-Correa 4 ,; Alejandro Coucheiro-Limeres 4 ,; Julia Olcoz 5 and; Antonio Miguel 5. EURASIP Journal on Audio, Speech … approach [6, 41–45] that searches for terms in the output of a large vocabulary continuous speech recognition (LVCSR) system, and the … Cited by 5 Related articles All 7 versions

Real-time translation of discrete Sinhala speech to Unicode text MKH Gunasekara… – Advances in ICT for …, 2015 – ieeexplore.ieee.org … Understanding, 2005. [27] Akinobu Lee, Kiyohiro Shikano, and Tatsuya Kawahara, “Recent progress of open-source LVCSR engine Julius and Japanese model repository,” 2004. [28] Thilini Nadungodage and Ruvan Weerasinghe … Related articles All 4 versions

Command-based voice teleoperation of a mobile robot via a human-robot interface A Poncela, L Gallardo-Estrella – Robotica, 2015 – Cambridge Univ Press … Thus, Julius/Julian was chosen as it is the voice decoder which better suited our requirements. Julius/Julian is an open-source large vocabulary continuous speech recognition (LVCSR) engine. It combines high recognition rate and high-speed speech recognition. … Cited by 3 Related articles All 5 versions

Deep belief neural networks and bidirectional long-short term memory hybrid for speech recognition ? Brocki, K Marasek – Archives of Acoustics, 2015 – degruyter.com … Trained acoustic models were incorporated to al- ready working LVCSR system (Korzinek, 2011). A standard system based on GMM models was also trained in HTK toolkit (Young, 2000) and tested in the Julius (Lee, 2001) decoder using a language model created by IRSTLM … Cited by 2 Related articles All 8 versions

Evaluation of optimal and sub-optimal speech noise reduction wiener filters ÍA Lima, MS Alencar, WTA Lopes… – … IWT), 2015 International …, 2015 – ieeexplore.ieee.org … Otherwise, the vector remains unaltered. The Large Vocabulary Continuous Speech Recognition (LVCSR) decoder Julius [12] is chosen for the ASR system. Since Julius is language-independent decoder, it is necessary to provide language and an acoustic models. … Related articles

Voice over Internet protocol: A collaborative tool to advance communication processes for elderly people M Zamikhaya, JS Van Der Walt… – … Conference (SAI), 2015, 2015 – ieeexplore.ieee.org … Julius provides two pass Large Vocabulary Speech Recognition (LVCSR) decoder software for speech related researchers and developers and it is based on word N-gram and context dependent Hidden Markov Models [10]. … Related articles All 3 versions

Integrated Interfacing System for Video Laparoscopy Procedures Based on Voice Recognition LF Aguiar, L Magalhães, RH Matsunaga… – VI Latin American …, 2015 – Springer … The voice recognition decoder used in this work is Jul- ian, which is a derived version of Julius [6]. It is an open source software for Large Vocabulary Continuous Speech Recognition (LVCSR) in real-time. … 6. Julian at http://www.voxforge.org/home/docs/julius- manual/julian. … Related articles

Situated language understanding for a spoken dialog system within vehicles T Misu, A Raux, R Gupta, I Lane – Computer Speech & Language, 2015 – Elsevier In this paper, we address issues in situated language understanding in a moving car, which has the additional challenge of being a rapidly changing environment. Cited by 4 Related articles All 3 versions

Multimodal Interaction with Co-located Drones for Search and Rescue J Cacace, A Finzi, V Lippiello – ceur-ws.org … then combined. Speech recognition is based on Julius[12], a two-pass large vocabulary continuous speech recognition (LVCSR) engine. The proposed gesture recogni- tion system exploits the Thalmic Myo Armbend. This device … Cited by 1 Related articles All 2 versions

Sound source separation for robot audition using deep learning K Noda, N Hashimoto, K Nakadai… – … (Humanoids), 2015 IEEE …, 2015 – ieeexplore.ieee.org … sound features were processed with a Japanese automatic speech recognizer, Julius [25]. VI. RESULTS AND DISCUSSION Fig. 6 and Fig. 7 summarize the WCRs over different noise levels and sound feature acquisition methods for the IWR and LVCSR tasks, respectively. … Related articles

Reducing morpho-phonetic confusion in sub-word based Uyghur ASR M Ablimit, A Hamdulla, A Pattar – Signal and Information …, 2015 – ieeexplore.ieee.org … [3] Lee, T. Kawahara, and K. Shikano, “Julius — an open source real-time large vocabulary recognition engine,” In Proc. … 27, June. 2001 [7] M. Jongtaveesataporn, I. Thienlikit, C. Wutiwiwatchai, S. Furui, “Lexical units for Thai LVCSR,” Speech Communication, pp.379–389, … Related articles

Recognition of OOV proper names in diachronic audio news I Sheikh, I Illina, D Fohr – Information Systems and Economic …, 2015 – ieeexplore.ieee.org … The ANTS [17] LVCSR system is used to perform auto- matic segmentation and speech-to-text transcription of the test audio news. ANTS is based on Context Dependent HMM phone models trained on 200 hour broadcast news audio files. ANTS uses the Julius [24] speech … Related articles All 2 versions

Singer Module with Singing Synthesis for Humanoid Robots Applications A Rojo-Hernandez, HM Perez-Meana… – Proceedings on the …, 2015 – search.proquest.com … 5.1 Julius. Julius is a high-performance, two pass large vocabulary continuous speech recognition (LVCSR) decoder software for speech-related researchers and developers. It can perform almost real-time decoding on most … Related articles

Methods for rapid development of automatic speech recognition system for Russian R Safarik, J Nouza – Electronics, Control, Measurement, Signals …, 2015 – ieeexplore.ieee.org … well summarized in the recently (2014) published journal paper [7]. Its authors describe an LVCSR system for Russian designed with the use of some open source tools (namely the HTK for AM training, the SRILM for LM training and Julius decoder for recognition experiments). … Related articles All 2 versions

Modeling of Slovak Language for Broadcast News Transcription J Stas, J Juhár – Journal of Electrical and Electronics …, 2015 – search.proquest.com … Available: http://www.inesc-id.pt/pt/indicadores/Ficheiros/3688.pdf. [14] J. Stas, D. Hládek and J. Juhár, ”Language model adaptation for Slovak LVCSR”, in Proc. … [16] A. Lee, T. Kawahara and K. Shikano, ”Julius – An open source real-time large vocabulary recognition engine”, in. … Related articles All 4 versions

Trinity Robotics Application for RoCKIN 2015 C McGinn, M Cullinan, M Culleton, A Sena, C Donovan… – rockincompetition.eu … Speech recognition function- ality is presently being implemented using a ROS package which integrates the Julius large vocabulary continuous speech recognition (LVCSR) decoder software [8]. However the group is also exploring using packages which facilitate both local … Related articles All 2 versions

Topic identification techniques applied to dynamic language model adaptation for automatic speech recognition JD Echeverry-Correa, J Ferreiros-López… – Expert Systems with …, 2015 – Elsevier … in spoken dialogue systems for adapting the speech recognizer to the dialogue context (López-Cózar and Callejas, 2006 and Lucas-Cuesta, 2013), in dynamic language model adaptation for Large Vocabulary Continuous Speech Recognition Systems (LVCSR) (Gollan … Cited by 4 Related articles All 4 versions

Neural Networks Revisited for Proper Name Retrieval from Diachronic Documents I Illina, D Fohr – LTC Language & Technology Conference, 2015 – hal.archives-ouvertes.fr … By searching new proper names and by adding them to the standard vocabulary of LVCSR, we want to face the problem of out-of-vocabulary words (OOV, words that are not in LVCSR system vocabulary). … The recognition engine is Julius (Lee and Kawahara, 2009). … Related articles All 2 versions

Improved Hindi Broadcast ASR by Adapting the Language Model and Pronunciation Model Using A Priori Syntactic and Morphophonemic Knowledge P Jyothi, M Hasegawa-Johnson – Sixteenth Annual Conference of …, 2015 – isle.illinois.edu … 703–715, 2004. [2] M. Rajat, Babita, and K. Abhishek, “Domain specific speaker in- dependent continuous speech recognition using Julius,” in Proc. … of PROPOR, 2003. [17] C. Huang, Y. Shi et al., “Segmental tonal modeling for phone set design in Mandarin LVCSR,” in Proc. … Related articles All 7 versions

A Comparison of RNN LM and FLM for Russian Speech Recognition I Kipyatkova, A Karpov – International Conference on Speech and …, 2015 – Springer … Speech Commun. 56, 213–228 (2014)CrossRef. 7. Kipyatkova, I., Karpov, A.: Lexicon size and language model order optimization for Russian LVCSR. … 234–246 (2015). 9. Lee, A., Kawahara, T.: Recent development of open-source speech recognition engine Julius. … Cited by 2 Related articles All 2 versions

Feature extraction using pre-trained convolutive bottleneck nets for dysarthric speech recognition Y Takashima, T Nakashika… – … (EUSIPCO), 2015 23rd …, 2015 – ieeexplore.ieee.org … [9] K. Vesely et al., “Convolutive bottleneck network fea- tures for LVCSR,” in ASRU, 2011, pp. 42–47. … [16] “Open-Source Speech Recognition Software Julius,” http://julius.sourceforge.jp/. 2015 23rd European Signal Processing Conference (EUSIPCO) 1415 Cited by 1 Related articles All 4 versions

Phonetic Segmentation Using KALDI and Reduced Pronunciation Detection in Causal Czech Speech Z Patc, P Mizera, P Pollak – International Conference on Text, Speech, and …, 2015 – Springer … on HTK and allows graphical user interfaces (GUI) using Praat environment [6], the SPPAS tool is based on Julius Speech Recognition … state transduc- ers (WFSTs) which are advantageous mainly for the purpose of Large Vocabu- lary Continuous Speech Recognition (LVCSR). … Cited by 2 Related articles All 3 versions

SAPTE: A multimedia information system to support the discourse analysis and information retrieval of television programs MHR Pereira, CL de Souza, FLC Pádua… – Multimedia Tools and …, 2015 – Springer … the videos, by considering aspects, such as, definitions of the subject of analysis, the nature of the speaker and the corpus of data resulting from the discourse; (2) a state of the art decoder software for large vocabulary continuous speech recognition, called Julius; (3) image and … Cited by 6 Related articles All 5 versions

Automatic speech recognition based Odia system B Karan, J Sahoo, PK Sahu – Microwave, Optical and …, 2015 – ieeexplore.ieee.org … There are many open source speech recognition toolkits are available like HTK, Julius, CMU and Sphinx. … Our proposed system can be implemented for LVCSR (Large vocabulary continuous speech recognition). ACKNOWLEDGMENTS …

Two-Step Spoken Term Detection using SVM Classifier Trained with Pre-Indexed Keywords based on ASR Result K Domoto, T Utsuro, N Sawada… – … Annual Conference of …, 2015 – researchgate.net … note-taking system [1]. STD is difficult to use when searching for terms within a vocabulary-free framework because search terms are not known by the STD process prior to the implementation of a large vocabulary continuous speech recognition (LVCSR) sys- tem. … Julius ver. … Cited by 1 Related articles All 3 versions

Nonparametric Bayesian double articulation analyzer for direct language acquisition from continuous speech signals T Taniguchi, S Nagasaka, R Nakashima – 2015 – ieeexplore.ieee.org Page 1. 2379-8920 (c) 2015 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/ redistribution requires IEEE permission. See http://www.ieee.org … Cited by 6 Related articles All 5 versions

Automatic Lecture Transcription Based on Discriminative Data Selection for Lightly Supervised Acoustic Model Training LI Sheng, Y Akita, T Kawahara – IEICE TRANSACTIONS on …, 2015 – search.ieice.org … [16] A. Lee and T. Kawahara, “Recent development of open-source speech recognition engine Julius,” Proc. … From 2009 to 2012, he served for CAS (Chinese Academic of Sciences), doing research on LVCSR, CALL and multimodal speech syn- thesis. … Related articles All 4 versions

Geo-location for Voice Search Language Modeling C Chelba, X Zhang, K Hall – Interspeech, to appear, 2015 – research.google.com … 5260–5263. [7] Alex Acero, Neal Bernstein, Rob Chambers, Yun-Cheng Ju, Xiao Li, Julian Odell, Patrick … 4, pp. 1352–1365, 2007. [13] George Saon, Daniel Povey, and Geoffrey Zweig, “Anatomy of an extremely fast LVCSR decoder,” in in Proc. Interspeech, 2005, pp. 549–552. Cited by 4 Related articles All 3 versions

Structured discriminative models using deep neural-network features RC van Dalen, J Yang, H Wang, A Ragni… – … IEEE Workshop on …, 2015 – ieeexplore.ieee.org Page 1. STRUCTURED DISCRIMINATIVE MODELS USING DEEP NEURAL-NETWORK FEATURES RC van Dalen, J. Yang, H. Wang, A. Ragni, C. Zhang, MJF Gales Department of Engineering, University of Cambridge, United Kingdom … Cited by 2 Related articles All 5 versions

Recurrent Neural Network-based Language Modeling for an Automatic Russian Speech Recognition System I Kipyatkova, A Karpov – Artificial Intelligence and Natural Language and …, 2015 – fruct.org … [18] I. Kipyatkova, A. Karpov, “Lexicon Size and Language Model Order Optimization for Russian LVCSR”, Springer International Publishing Switzerland. … 3161 3164. [26] A. Lee, T. Kawahara, “Recent Development of Open-Source Speech Recognition Engine Julius”, in Proc. … Cited by 1 Related articles All 2 versions

Automatic classification of usability of ASR result for real-time captioning of lectures Y Akita, N Kuwahara… – 2015 Asia-Pacific Signal …, 2015 – ieeexplore.ieee.org … Both systems used the same ASR setting. As a decoder, we used Julius rev. … Speech & Audio Process., Vol.9, No.3, pp.288–298, 2001. [4] Z.Zhou, H.Meng, and WKLo, “A Multi-pass Error Detection and Correc- tion Framework for Mandarin LVCSR,” In Proc. … Related articles All 2 versions

Optimized wavelet-domain filtering under noisy and reverberant conditions R Gomez, T Kawahara… – APSIPA Transactions on …, 2015 – Cambridge Univ Press Page 1. SIP (2015), vol. 4, e3, page 1 of 12 © The Authors, 2015. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/3.0/), which permits … Cited by 1 Related articles

A 6 mW, 5,000-Word Real-Time Speech Recognizer Using WFST Models M Price, J Glass… – IEEE Journal of Solid-State …, 2015 – ieeexplore.ieee.org … 16]–[18] bridged the gap between high-performance and low-power applications, ap- plying a series of optimizations to a Japanese-language system called Julius in order … [19] JR Novak, N. Minematsu, and K. Hirose, “Open source WFST tools for LVCSR cascade development … Cited by 4 Related articles All 7 versions

Thinking Required K Rocki – arXiv preprint arXiv:1512.01926, 2015 – arxiv.org Page 1. Thinking Required Kamil Rocki kmrocki@us.ibm.com IBM Research, San Jose, CA 95120, USA Abstract There exists a theory of a single general-purpose learn- ing algorithm which could explain the principles its operation. … Related articles All 4 versions

Culture Clubs Processing Speech by Deriving and Exploiting Linguistic Subcultures DG Brizan – 2015 – gc.cuny.edu … speech, conversational and spontaneous speech and others) [77] and vocabulary size (small through large). My work on the corpora listed in Table 1 will largely be considered LVCSR (large vocabulary, continuous speech recognition). … Related articles All 2 versions

A Study on Acoustic Modeling of Speech for Personalized Speech Interface Y Ijima – A Study on Acoustic Modeling of Speech for …, 2015 – t2r2.star.titech.ac.jp Page 1. ?? / ???? Article / Book Information ??(??) Title(English) A Study on Acoustic Modeling of Speech for Personalized Speech Interface ??(??) ???? Author(English) Yusuke Ijima ??(??) ??:????? … Related articles All 2 versions

Graph-based Semi-supervised Acoustic Modeling for Automatic Speech Recognition Y Liu – 2015 – ssli.ee.washington.edu Page 1. Graph-based Semi-supervised Acoustic Modeling for Automatic Speech Recognition Yuzong Liu Mar 30, 2015 Abstract In this thesis proposal, we investigate how to apply graph-based semi-supervised learning to acoustic modeling in speech recognition. … Related articles All 2 versions