%Aigaion2 BibTeX export from Idiap Publications
%Thursday 21 November 2024 11:33:18 AM

@INPROCEEDINGS{Hermann_INTERSPEECH_2023,
         author = {Hermann, Enno and Magimai.-Doss, Mathew},
       keywords = {Automatic Speech Recognition, Dysarthria, Few-shot learning, speech synthesis},
       projects = {Idiap, TAPAS, IICT},
          title = {Few-shot Dysarthric Speech Recognition with Text-to-Speech Data Augmentation},
      booktitle = {Proceedings of Interspeech},
           year = {2023},
          pages = {156-160},
            url = {https://www.isca-speech.org/archive/interspeech_2023/hermann23_interspeech.html},
            doi = {10.21437/Interspeech.2023-2481},
       abstract = {Speakers with dysarthria could particularly benefit from assistive speech technology, but are underserved by current automatic speech recognition (ASR) systems. The differences of dysarthric speech pose challenges, while recording large amounts of training data can be exhausting for patients. In this paper, we synthesise dysarthric speech with a FastSpeech 2-based multi-speaker text-to-speech (TTS) system for ASR data augmentation. We evaluate its few-shot capability by generating dysarthric speech with as few as 5 words from an unseen target speaker and then using it to train speaker-dependent ASR systems. The results indicated that, while the TTS output is not yet of sufficient quality, this could allow easy development of personalised acoustic models for new dysarthric speakers and domains in the future.},
            pdf = {https://publications.idiap.ch/attachments/papers/2023/Hermann_INTERSPEECH_2023.pdf}
}