Institut

Studium

Forschung


 

   AUTOMATIC SEGMENTATION
OF
SPOKEN
LANGUAGE

  The ALIGN Project  Stuttgart - Mannheim

   
  Segmentation of Speech Data from the Mannheim Institute of German Language

Maschinelle Segmentierung von Korpora gesprochener Sprache des IDS (Alignment-Projekt)

The aim of this project is to transform the Speech Data into small units using Automatic Speech Recognition (ASR) methods. Such a tranformation is traditionally called a transcription of speech signals. The transcription is elementary of every machine-based access to data. The automatic segmentation is archieved by ASR-Tools (a so-called  Aligner) developed by Stefan Rapp.

The project is divided into three stages. In the first stage a few examples of the speech data are tested, the results are evaluated and the errors are classified. In the second stage the errors are systematically investigated, allowing their reestimations. In the third stage the speech data were processed by means of the reestimated Aligner.

The project is supported by the Ministry of Education, Research and Cultre of Baden- Wuertemberg .


   
  RESEARCH PROGRAMME
STAFF
LABS & EQUIPMENT
ACTIVITIES


maintained by J. Mayer,  Marcus Fach
© 1997,1998