Interactive Systems Labs (ISL)

Janus Recognition Toolkit

The Janus Recognition Toolkit (JRTk) is a general-purpose speech recognition toolkit developed at the Interactive Systems Labs at Carnegie Mellon University and Karlsruhe Institute of Technology. Commercial and research liscenses are available.

The Janus Recognition Toolkit (JRTk) is a general-purpose speech recognition toolkit useful for both research and application development and is part of the JANUS speech-to-speech translation system.

The JRTk provides a flexible Tcl/Tk script based environment which enables researchers to build state-of-the-art speech recognizers and allows them to develop, implement, and evaluate new methods. It implements an object oriented approach that unlike other toolkits is not a set of libraries and precompiled modules but a programmable shell with transparent, yet efficient objects.

Since version 5 JRTk features the IBIS decoder, a one-pass decoder that is based on a re-entrant single pronunciation prefix tree and makes use of the concept of linguistic context polymorphism. It is therefore able to incorporate full linguistic knowledge at an early stage. It is possible to decode in one pass, using the same engine in combination with a statistical n-gram language model as well as context- free grammars. It is also possible to use the decoder to rescore lattices in a very efficient way.

JRTk features state-of-the art techniques for pre-processing, acoustic modeling, and search.

Acoustic Pre-Processing

  • Processing of various, frequent audio formats
  • Flexible short-term fourier analysis
  • Flexibly configurable Mel-frequency scaled cepstral coefficients calculation
  • Minimimum variance distortion response processing
  • LPC processing
  • Mean and variance normalization

Acoustic Modeling

  • EM Training, label training, Viterbi training
  • incremental growing of Gaussians
  • Semi-tied covariances
  • MMIE training, bMMIE training
  • Speaker adaptive training

 

 

Decoding

  • Single pass decoder
  • Flexible language model interface for n-gram language models and grammars
  • lattice generation and manipulation
  • lattice rescoring
  • consensus decoding
  • confusion network combination
The JRTk is used for speech recognition in many on-going projects as well as past ones.

JRTk in current projects

Lecture Translator The ISL’s Simultaneous Lecture Translation system is capable of automatically translating lectures and presenting the translation results as text via the www. The system is used to translate German lectures into English and further languages, so that international students can better follow their content, even if they are not fluent in German.  

 

JRTk in past projects

EU-BRIDGE The project will provide streaming technology that can convert speech from lectures, meetings, and telephone conversations into the text in another language. ISL is coordinator of the project.
EVEIL-3D Im Projekt EVEIL-3D wurde ein Serious Game für den Einsatz im Rahmen des schulischen Fremdsprachenunterrichts entwickelt.
TC-STAR TC-STAR is envisioned as a long term effort focused on advanced research in all core technologies for speech to speech translation (SST): speech recognition, speech translation and speech synthesis.
C-STAR

The Consortium for Speech Translation Advanced Research (C-STAR) has emerged from originally informal bilateral collaborations between research labs interested in Automatic Translation of Spoken Language.

FAME
Facilitating Agents for Multi-Cultural Exchange - FAME
The vision of the FAME project was to construct and intelligent agent to facilitate communication among people from different cultures who collaborate on solving a common problem.
View4You View 4 You automatically records the "Tagesschau" every day and allows the user to retrieve video segments of news items for different topics using spoken language input.
CHIL CHIL - Computers in the Human Interaction Loop - is an Integrated Project (IP 506909) under the European Commission's Sixth Framework Programme. I
PF-STAR  
NESPOLE! The project aims at supporting multilingual and multimodal negotiation in e-commerce and e-service by providing a robust, flexible, scalable and portable speech-to-speech translation system.
VERBMOBIL The Verbmobil system recognizes spontaneous speech, analyzes the input and translates it into a Foreign language, creates a sentence and pronounces it.
BABEL Rapid porting of keyword search for new languages
Quaero Quaero is a program promoting research and industrial innovation on technologies for automatic analysis and classification of multimedia and multilingual documents. 
SFB 588 Humanoid Robots The goal of this interdisciplinary research project is the development of humanoid robots which resemble humans in their ways of acting in the world, of reasoning and of communicating about the world.

License Information

The JRTk is liscensed by Carnegie Mellon University. Commercial as well as research liscenses are available. Terms and conditions as well as further information can be inquired by contacting Prof. Alex Waibel at ahw∂cs.cmu.edu.

 

JRTk Articles
Title Author Source

Bernhard Suhm, Petra Geutner, Thomas Kemp, Alon Lavie, Laura Tomokiyo, Arthur McNair, Ivica Rogina, Tanja Schultz, Tilo Sloboda, Wayne Ward, Monika Woszczyna, Alex Waibel 

01. January 1995 

Donna Gates, Alon Lavie, Lori Levin, Alex Waibel, Marsal Gavalda, Laura Tomokiyo, Monika Woszczyna, Puming Zhan 

12th European Conference on Artificial Intelligence, ECAI 1996, Budapest, Hungary, 01. August 1996 

Alon Lavie, Alex Waibel, Lori Levin, Donna Gates, Marsal Gavalda, Torsten Zeppenfeld, Puming Zhan, Oren Glickman 

4th International Conference on Spoken Language Processing 1996, ICSLP 1996, Philadelphia, USA, 01. October 1996 

Puming Zhan, Klaus Ries, Marsal Gavalda, Donna Gates, Alon Lavie, Alex Waibel 

4th International Conference on Spoken Language Processing 1996, ICSLP 1996, Philadelphia, USA, 01. October 1996 

Hagen Soltau, Florian Metze, Christian Fügen, Alex Waibel

Automatic Speech Recognition and Understanding Workshop 2001, ASRU 2001, Trento, Italy, 25. October 2011

Thomas Polzin, Noah Coccaro, N. Aoki-Waibel, Monika Woszczyna, M. Tomita, J. Tsutsumi, Ivica Rogina, Carolyn Rose, Alex Waibel, Arthur McNair, Alon Lavie, A. Eisele, Tilo Sloboda, Wayne Ward 

European Conference on Speech Communication and Technology 1993, Eurospeech 1993, Berlin, Germany, 26. January 1993 

Alex Waibel, Michael Finke, Donna Gates, Marsal Gavalda, Thomas Kemp, Alon Lavie, Lori Levin, Uwe Meier, Laura Tomokiyo, Arthur McNair, Ivica Rogina, Kaori Shima, Tilo Sloboda, Monika Woszczyna, Torsten Zeppenfeld, Puming Zhan 

IEEE International Conference On Acoustics, Speech And Signal Processing 1996, ICASSP 1996, Atlanta, USA, 01. May 1996 

Louise Osterholtz, Joe Tebelskis, Ivica Rogina, Hiroaki Saito, Charles Augustine, Arthur McNair, Alex Waibel, Monika Woszczyna, Tilo Sloboda 

IEEE International Conference on Acoustics, Speech, and Signal Processing 1992, ICASSP 1992, San Francisco, USA, 26. January 1992 

Monika Woszczyna, N. Aoki-Waibel, Finn Dag Buø, Noah Coccaro, Keiko Horiguchi, Thomas Kemp, Alon Lavie, Arthur McNair, Thomas Polzin, Ivica Rogina, Carolyn Rose, Tanja Schultz, Bernhard Suhm, M. Tomita, Alex Waibel 

International Conference on Acoustics, Speech, and Signal Processing 1994, ICASSP 1994, Adelaide, Australia, 01. April 1994 

Alon Lavie, Alex Waibel, Lori Levin, Michael Finke, Donna Gates, Marsal Gavalda, Torsten Zeppenfeld, Puming Zhan 

International Conference on Acoustics, Speech, and Signal Processing 1997, ICASSP 1997, Munich, Germany, 01. April 1997 

Torsten Zeppenfeld, Michael Finke, Klaus Ries, Martin Westphal, Alex Waibel 

International Conference on Acoustics, Speech, and Signal Processing 1997, ICASSP 1997, Munich, Germany, 01. April 1997