论文信息 - Phonemic Segmentation and Labelling using the MAUS Technique

Phonemic Segmentation and Labelling using the MAUS Technique

We describe the pronunciation model of the automatic segmentation technique MAUS based on a data-driven Markov process and a new evaluation measure for phonemic transcripts relative symmetric accuracy; results are given for the MAUS segmentation and labelling on German dialog speech. MAUS is currently distributed as a freeware package by the Bavarian Archive for Speech Signals and will also be implemented as a web-service in the near future.

Jonathan Harrington | Florian Schiel | Christoph Draxler

[1] Wolfgang Wahlster,et al. Verbmobil: Foundations of Speech-to-Speech Translation , 2000, Artificial Intelligence.

[2] SchielFlorian,et al. Alcohol language corpus , 2012, LREC 2012.

[3] Helmer Strik,et al. Improving the performance of a Dutch CSR by modeling pronunciation variation , 1998 .

[4] Florian Schiel,et al. Alcohol language corpus: the first public corpus of alcoholized German speech , 2012, Lang. Resour. Evaluation.

[5] Florian Schiel,et al. Automatic Phonetic Transcription of Non-Prompted Speech , 1999 .

[6] Christoph Draxler,et al. Wikispeech - a content management system for speech databases , 2008, INTERSPEECH.

[7] Andreas Kipp. Automatische Segmentierung und Etikettierung von Spontansprache , 1999 .

[8] Florian Schiel,et al. Verbmobil Data Collection and Annotation , 2000 .

[9] J. Fleiss. Measuring nominal scale agreement among many raters. , 1971 .

[10] Jacob Cohen. A Coefficient of Agreement for Nominal Scales , 1960 .

[11] Florian Schiel,et al. Signal processing via web services: The use case WebMAUS , 2012 .

[12] Florian Schiel,et al. Automatic detection and segmentation of pronunciation variants in German speech corpora , 1996, Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96.