Doc. Dr. Ing. Jan Černocký

Data driven and anthropic coding and recognition of speech
Reseach leader:Černocký Jan
Agency:GAČR
Code:GP102/02/D108
Start:2002
End:2005
Keywords:speech processing, coding, recognition, data-driven methods
Annotation:
Data driven and anthropic coding and recognition of speech
Project description:
The topic of the proposed project are two fields of speech processing: robust speech recognition and very low bit-rate coding. The principal problem of both fields is the choice of processing methods and of basic speech units. This choice is often given historically, without a thorough investigation of the nature of actually processed speech data. The proposed project is aimed to two directions to overcome this negative trend: to the development of methods and and to the discovery of speech units on the data (with a minimum of a-priori choices that must be done by the system designer) and to anthropic methods, that reflect to the greatest possible extent the speech processing by human auditory system. In the field of selection of units, we will concentrate on ALISP (Automatic language independent speech processing) methods, that limit the necessity of manual transcription of speech databases. In the field of robust recognition, we will use methods based on linear discriminant analysis (LDA) and on neural networks. Actual outputs of the project include a speech coder working on the top of TCP/IP protocol with a bit rate of several hundreds of bps and a demonstrator of on-line telephone speech recognizer with robust parameterization.

Related projects

2002Voice technologies for support of information society, GAČR, GA102/02/0124, 2002-2004, completed
Research leader: Černocký Jan
Team leaders: Burget Lukáš, Grézl František, Karafiát Martin, Motlíček Petr, Schwarz Petr

Publications

2005Černocký Jan, Lampa Petr: Teaching signals - making it automatic, making it fun, In: Proc. Radioelektronika 2005, Brno, CZ, FEKT VUT, 2005, p. 4, ISBN 80-214-2904-6
 Matějka Pavel, Schwarz Petr, Černocký Jan, Chytil Pavel: Phonotactic Language Identification using High Quality Phoneme Recognition, In: Interspeech'2005 - Eurospeech - 9th European Conference on Speech Communication and Technology, Lisbon, PT, ISCA, 2005, p. 2237-2240, ISSN 1018-4074
 Motlíček Petr, Burget Lukáš, Černocký Jan: VISUAL FEATURES FOR MULTIMODAL SPEECH RECOGNITION, In: Radioelektronika 2005, Brno, CZ, FEKT VUT, 2005, p. 187-190, ISBN 80-214-2904-6
 Szőke Igor, Schwarz Petr, Burget Lukáš, Fapšo Michal, Karafiát Martin, Černocký Jan, Matějka Pavel: Comparison of Keyword Spotting Approaches for Informal Continuous Speech, In: Interspeech'2005 - Eurospeech - 9th European Conference on Speech Communication and Technology, Lisabon, PT, 2005, p. 633-636, ISSN 1018-4074
 Szőke Igor, Schwarz Petr, Burget Lukáš, Karafiát Martin, Černocký Jan: Phoneme based acoustics keyword spotting in informal continuous speech, In: Radioelektronika 2005, Brno, CZ, FEKT VUT, 2005, p. 195-198, ISBN 80-214-2904-6
 Szőke Igor, Schwarz Petr, Burget Lukáš, Karafiát Martin, Matějka Pavel, Černocký Jan: Phoneme Based Acoustics Keyword Spotting in Informal Continuous Speech, In: Lecture Notes in Computer Science, Vol. 2005, No. 3658, DE, p. 8, ISSN 0302-9743
 Szőke Igor, Schwarz Petr, Matějka Pavel, Burget Lukáš, Fapšo Michal, Karafiát Martin, Černocký Jan: Comparison of Keyword Spotting Approaches for Informal Continuous Speech, In: 2nd Joint Workshop on Multimodal Interaction and Related Machine Learning Algorithms, Edinburgh, GB, 2005, p. 12
 Szőke Igor: Smooth Pitch Tracker Based on Harmonic and Noise Model, In: STUDENT EEICT 2005, Brno, CZ, FIT VUT, 2005, p. 673-677, ISBN 80-214-2890-2
2004Karafiát Martin, Grézl František, Černocký Jan: TRAP based features for LVCSR of meeting data, In: Proc. 8th International Conference on Spoken Language Processing, Jeju Island, KR, Sunjin, 2004, p. 437-440, ISSN 1225-4111
 Matějka Pavel, Černocký Jan, Sigmund Milan: Introduction to Automatic Language Identification, In: Conference Proceedings of Radioelektronika 2004, Brno, CZ, STUBA, 2004, p. 4, ISBN 80-227-2017-8
 Matějka Pavel, Szőke Igor, Schwarz Petr, Černocký Jan: Automatic Language Identification using Phoneme and Automatically Derived Unit Strings, In: Proceedings of 7th International Conference Text,Speech and Dialoque 2004, Brno, CZ, Springer, 2004, p. 8, ISBN 3-540-23049-1
 Matějka Pavel, Szőke Igor, Schwarz Petr, Černocký Jan: Automatic Language Identification using Phoneme and Automatically Derived Unit Strings, In: Lecture Notes in Computer Science, Vol. 2004, No. 3206, DE, p. 8, ISSN 0302-9743
 Motlíček Petr, Burget Lukáš, Černocký Jan: PHONEME RECOGNITION OF MEETINGS USING AUDIO-VISUAL DATA, AMI Workshop, Martigny, CH, 2004, p. 6
 Motlíček Petr, Černocký Jan: Multimodal Phoneme Recognition of Meeting Data, In: 7th International Conference, TSD 2004 Brno, Czech Republic, September 2004 Proceedings, Brno, CZ, Springer, 2004, p. 379-384, ISBN 3-540-23049-1, ISSN 0302-9743
 Motlíček Petr, Černocký Jan: Multimodal Phoneme Recognition of Meeting Data, In: Lecture Notes in Computer Science, Vol. 2004, No. 3206, DE, p. 6, ISSN 0302-9743
 Schwarz Petr, Matějka Pavel, Černocký Jan: Phoneme Recognition from a Long Temporal Context, In: poster at JOINT AMI/PASCAL/IM2/M4 Workshop on Multimodal Interaction and Related Machine Learning Algorithms, Martigny, CH, IDIAP, 2004, p. 1-1
 Schwarz Petr, Matějka Pavel, Černocký Jan: Towards Lower Error Rates in Phoneme Recognition, In: Proceedings of 7th International Conference Text,Speech and Dialoque 2004, Brno, CZ, Springer, 2004, p. 8, ISBN 3-540-23049-1
 Schwarz Petr, Matějka Pavel, Černocký Jan: Towards Lower Error Rates in Phoneme Recognition, In: Lecture Notes in Computer Science, Vol. 2004, No. 3206, DE, p. 8, ISSN 0302-9743
2003Burget Lukáš, Černocký Jan: Recognition of Speech with Non-random Attributes, In: 6th International Conference, TSD 2003 České Budějovice, Czech Republic, September 2003 Proceedings, České Budějovice, CZ, Springer, 2003, p. 6, ISBN 3-540-20024-X, ISSN 0302-9743
 Černocký Jan: Temporal processing for feature extraction in speech recognition, Vědecké spisy VUT, Brno, CZ, VUTIUM, 2003, p. 1-30, ISBN 80-214-2395-1
 Heřmanský Hynek, Matějka Pavel, Schwarz Petr: On Use of Temporal Dynamics of Speech for Language Identification, In: Proceedings of Language Recognition Workshop 2003, NIST Gaithersburg, MD USA, US, 2003, p. 56-62
 Matějka Pavel, Schwarz Petr, Grézl František, Černocký Jan: Phoneme Classification using Temporal Patterns, In: Proc. 13th International scientific conference Radioelektronika 2003, Brno, CZ, FEKT VUT, 2003, p. 1-4, ISBN 80-214-2383-8
 Matějka Pavel, Schwarz Petr, Heřmanský Hynek, Černocký Jan: Phoneme Recognition using Temporal Patterns, In: Proc. 6th International Conference Text, Speech and Dialogue, TSD2003, Ceske Budejovice, CZ, Springer, 2003, p. 465-472, ISBN 3-540-20024-X
 Motlíček Petr, Černocký Jan: All-Pole Modeling for Definition of Speech Features in Aurora3 DSR Task, In: 6th International Conference, TSD 2003 České Budějovice, Czech Republic, September 2003 Proceedings, České Budějovice, CZ, ZČU v Plzni, 2003, p. 295-300, ISBN 3-540-20024-X, ISSN 0302-9743
 Motlíček Petr, Černocký Jan: Autoregressive Modeling based Feature Extraction for Aurora3 DSR Task, In: Proc. EUROSPEECH 2003, Geneva, CH, IDIAP, 2003, p. 1801-1804, ISSN 1018-4074
 Motlíček Petr, Černocký Jan: Time-domain based Temporal Processing with Application of, In: Proc. EUROSPEECH 2003, Geneva, CH, IDIAP, 2003, p. 821-824, ISSN 1018-4074
 Schwarz Petr, Matějka Pavel, Černocký Jan: Recognition of Phoneme Strings using TRAP Technique, In: Proceedings of 8th International Conference Eurospeech, Geneve, CH, ISCA, 2003, p. 1-4, ISSN 1018-4074
 Schwarz Petr: Would You Like To Make Your Programs Understand Human Voice?, In: Proceedings of 9th Conference STUDENT EEICT 2003, Brno, CZ, FEKT VUT, 2003, p. 231-235, ISBN 80-214-2379-X
2002Černocký Jan: Temporal processing for feature extraction in speech recognition, habilitation thesis, Brno, CZ, 2002, p. 80
 Motlíček Petr, Burget Lukáš: Efficient Noise Estimation and its Application for Robust Speech Recognition, In: 5th International Conference, TSD 2002 Brno, Czech Republic, September 2002 Proceedings, Berlin, DE, Springer, 2002, p. 229-236, ISBN 3-540-44129-8
 Motlíček Petr: Application of Mel-scale Filter bank for Noise Estimation in Speech Processing, In: 12th International Czech-Slovak Scientific conference Radioelektronika 2002, Bratislava, SK, STUBA, 2002, p. 4, ISBN 80-227-1700-2
 Motlíček Petr: Feature Extraction in Speech Coding and Recognition, Portland, US, OGI, 2002, p. 1-50

Your IPv4 address: 23.22.76.170
Switch to IPv6 connection

DNSSEC [dnssec]