A System for Information Retrieval from Large Records of Czech Spoken Data

  • Jan Nouza
  • Jindřich Žďánský
  • Petr Červa
  • Jan Kolorenč
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4188)


In the paper we describe a complex multi-level system that serves for automatic search in large records of Czech spoken data. It includes modules for audio signal segmentation, speaker identification and adaptation, speech recognition and full-text search. The search can focus both on key-words and key-speakers. The transcription accuracy is about 79 % (for broadcast programs), search accuracy about 90 %. Due to its distributed platform, the system can operate in almost real-time.


Speech Recognition Audio Signal Automatic Speech Recognition Speaker Identification Broadcast News 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Semantic Retrieval of Multimedia. IEEE Signal Processing Magazine (March 2006)Google Scholar
  2. 2.
    Makhoul, J., et al.: Speech and Language Technologies for Audio Indexing and Retrieval. Proc. IEEE 88(8), 1338–1353 (2000)CrossRefGoogle Scholar
  3. 3.
    Zhou, B., Hansen, J.H.L.: Speechfind: An Experimental On-line Spoken Document Retrieval System for Historical Audio Archives. In: Proc. of ICSLP 2002, Denver (2002)Google Scholar
  4. 4.
    Byrne, W., et al.: Aut. recognition of spontaneous speech for access to multilingual oral history archives. IEEE Trans. on SAP 12(4), 420–435Google Scholar
  5. 5.
    Favre, B., Bechet, F., Nocera, P.: Mining Broadcast News Data: Robust Information Extraction from Words Lattices. In: Proc. of EuroSpeech 2005, Lisbon (September 2005)Google Scholar
  6. 6.
    Kurimo, M., Turunen, V., Ekman, I.: An Evaluation of a Spoken Document Retrieval Baseline System in Finnish. In: Proc. of ICSLSP 2004, Jeju (October 2004)Google Scholar
  7. 7.
    Ždánský, J.: Methods for Speaker Detection Change Identification in Audio Signal. PhD thesis (in Czech). Technical University in Liberec (October 2005)Google Scholar
  8. 8.
    Červa, P., David, P., Nouza, J.: Acoustic Modeling Based on Speaker Recognition and Adaptation for Improved Transcription of Broadcast Programs. In: Proc. of Specom 2005, Patras, Greece, pp. 183–186 (October 2005)Google Scholar
  9. 9.
    Nouza, J., Ždánský, J., David, P., Červa, P., Kolorenc, J., Nejedlova, D.: Fully Automated System for Czech Spoken Broadcast Transcription with Very Large (300K+) Lexicon. In: Proc. of Interspeech 2005, Lisboa, Portugal, pp. 1681–1684 (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Jan Nouza
    • 1
  • Jindřich Žďánský
    • 1
  • Petr Červa
    • 1
  • Jan Kolorenč
    • 1
  1. 1.SpeechLabTechnical University of LiberecLiberec 1Czech Republic

Personalised recommendations