Janus Recognition Toolkit

from Wikipedia, the free encyclopedia
Janus Recognition Toolkit
Basic data

Maintainer Christian Fügen and Florian Metze
Current  version V5.0 P014
programming language C, Tcl

The Janus Recognition Toolkit (short: JRTk) is a speech recognition toolkit that is being developed by the Interactive Systems Laboratories of Carnegie Mellon University and the Karlsruhe Institute of Technology . It can be used for research and application development and is part of the JANUS speech-to-speech translation system.

Janus offers Tcl / Tk programming interfaces. It was implemented in an object-oriented manner and is not a bunch of precompiled libraries, but a programmable shell.

From version 5 the JRTk includes the Ibis decoder. He is able to incorporate linguistic knowledge into the translation at an early stage. It is possible to decode with it in one go. A statistical n-gram language model and context-free grammars are used for this purpose . However, the generation of language models is not part of the JRTk. The Ibis decoder can, however, read in language models in the standardized ARPA Language Model File Format.

The JRTk uses hidden Markov models for the acoustic model and provides many techniques for acoustic preprocessing, for training the acoustic model, and for decoding speech.

The JRTk was used by the Interactive System Labs for various projects:

  • EU BRIDGE
  • EVEIl-3D
  • BABEL
  • Quaero
  • SFB 588
  • TC-STAR
  • FAME
  • Verbmobil
  • NESPOLE!

Web links

Individual evidence

  1. ↑ a b c d e f g h Janus Recognition Toolkit. Retrieved October 23, 2013 .
  2. ↑ Lori Lamel, Sandrine Courcinous, Julien Despres, Jean-Luc Gauvain, Yvan Josse, Kevin Kilgour, Florian Kraft2, Viet Bac Le; Hermann Ney, Markus Nußbaum-Thom, Ilya Oparin, Tim Schlippe, Ralf Schlüter, Tanja Schultz, Thiago Fraga da Silva, Sebastian Stüker, Martin Sundermeyer, Bianca Vieru, Ngoc Thang Vu, Alexander Waibel, Cecile Woehrling: Speech Recognition for Machine Translation in Quaero. (PDF; 239 kB) Retrieved October 23, 2013 .
  3. ↑ Michael Finke, Petra Geutner, Hermann Hild, Thomas Kemp, Klaus Ries, Martin Westphal: THE KARLSRUHE-VERBMOBIL SPEECH RECOGNITION ENGINE. (PDF; 134 kB) Retrieved October 23, 2013 .