International Workshop on Spoken Language Translation (IWSLT) 2011

San Francisco, CA, USA
December 8-9, 2011

The 2011 KIT QUAERO Speech-to-Text System for Spanish

Kevin Kilgour (1,2), Christian Saam (1,2), Christian Mohr (1), Sebastian Stüker (1,2), Alex Waibel (1)

(1) Institute of Anthropomatics; (2) Research Group 3-01 'Multilingual Speech Recognition'
Karlsruhe Institute of Technology, Karlsruhe, Germany

This paper describes our current Spanish speech-to-text (STT) system with which we participated in the 2011 Quaero STT evaluation that is being developed within the Quaero program. The system consists of 4 separate subsystems, as well as the standard MFCC and MVDR phoneme based subsystems we included a both a phoneme and grapheme based bottleneck subsystem. We carefully evaluate the performance of each subsystem. After including several new techniques we were able to reduce the WER by over 30% from 20.79% to 14.53%.

Full Paper

Bibliographic reference.  Kilgour, Kevin / Saam, Christian / Mohr, Christian / Stüker, Sebastian / Waibel, Alex (2011): "The 2011 KIT QUAERO speech-to-text system for Spanish", In IWSLT-2011, 199-205.