Domain-adversarial training for session independent EMG-based speech recognition

Michael Wand; Tanja Schultz; Jürgen Schmidhuber

doi:10.21437/Interspeech.2018-2318

Domain-adversarial training for session independent EMG-based speech recognition

Michael Wand, Tanja Schultz, Jürgen Schmidhuber

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

17 Scopus citations

Abstract

We present our research on continuous speech recognition based on Surface Electromyography (EMG), where speech information is captured by electrodes attached to the speaker's face. This method allows speech processing without requiring that an acoustic signal is present; however, reattachment of the EMG electrodes causes subtle changes in the recorded signal, which degrades the recognition accuracy and thus poses a major challenge for practical application of the system. Based on the growing body of recent work in domain-adversarial training of neural networks, we present a system which adapts the neural network frontend of our recognizer to data from a new recording session, without requiring supervised enrollment.

Original language	English (US)
Title of host publication	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Publisher	International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org
Pages	3167-3171
Number of pages	5
DOIs	https://doi.org/10.21437/Interspeech.2018-2318
State	Published - Jan 1 2018
Externally published	Yes

Bibliographical note

Generated from Scopus record by KAUST IRTS on 2022-09-14

Access to Document

10.21437/Interspeech.2018-2318

Cite this

Wand, M., Schultz, T., & Schmidhuber, J. (2018). Domain-adversarial training for session independent EMG-based speech recognition. In Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH (pp. 3167-3171). International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org. https://doi.org/10.21437/Interspeech.2018-2318

Wand, Michael ; Schultz, Tanja ; Schmidhuber, Jürgen. / Domain-adversarial training for session independent EMG-based speech recognition. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org, 2018. pp. 3167-3171

@inproceedings{110a70bfdd2549a8a3cf672aa92ab79f,

title = "Domain-adversarial training for session independent EMG-based speech recognition",

abstract = "We present our research on continuous speech recognition based on Surface Electromyography (EMG), where speech information is captured by electrodes attached to the speaker's face. This method allows speech processing without requiring that an acoustic signal is present; however, reattachment of the EMG electrodes causes subtle changes in the recorded signal, which degrades the recognition accuracy and thus poses a major challenge for practical application of the system. Based on the growing body of recent work in domain-adversarial training of neural networks, we present a system which adapts the neural network frontend of our recognizer to data from a new recording session, without requiring supervised enrollment.",

author = "Michael Wand and Tanja Schultz and J{\"u}rgen Schmidhuber",

note = "Generated from Scopus record by KAUST IRTS on 2022-09-14",

year = "2018",

month = jan,

day = "1",

doi = "10.21437/Interspeech.2018-2318",

language = "English (US)",

pages = "3167--3171",

booktitle = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

publisher = "International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org",

}

Wand, M, Schultz, T & Schmidhuber, J 2018, Domain-adversarial training for session independent EMG-based speech recognition. in Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org, pp. 3167-3171. https://doi.org/10.21437/Interspeech.2018-2318

Domain-adversarial training for session independent EMG-based speech recognition. / Wand, Michael; Schultz, Tanja; Schmidhuber, Jürgen.
Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org, 2018. p. 3167-3171.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Domain-adversarial training for session independent EMG-based speech recognition

AU - Wand, Michael

AU - Schultz, Tanja

AU - Schmidhuber, Jürgen

N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We present our research on continuous speech recognition based on Surface Electromyography (EMG), where speech information is captured by electrodes attached to the speaker's face. This method allows speech processing without requiring that an acoustic signal is present; however, reattachment of the EMG electrodes causes subtle changes in the recorded signal, which degrades the recognition accuracy and thus poses a major challenge for practical application of the system. Based on the growing body of recent work in domain-adversarial training of neural networks, we present a system which adapts the neural network frontend of our recognizer to data from a new recording session, without requiring supervised enrollment.

AB - We present our research on continuous speech recognition based on Surface Electromyography (EMG), where speech information is captured by electrodes attached to the speaker's face. This method allows speech processing without requiring that an acoustic signal is present; however, reattachment of the EMG electrodes causes subtle changes in the recorded signal, which degrades the recognition accuracy and thus poses a major challenge for practical application of the system. Based on the growing body of recent work in domain-adversarial training of neural networks, we present a system which adapts the neural network frontend of our recognizer to data from a new recording session, without requiring supervised enrollment.

UR - https://www.isca-speech.org/archive/interspeech_2018/wand18_interspeech.html

UR - http://www.scopus.com/inward/record.url?scp=85054987262&partnerID=8YFLogxK

U2 - 10.21437/Interspeech.2018-2318

DO - 10.21437/Interspeech.2018-2318

M3 - Conference contribution

SP - 3167

EP - 3171

BT - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

PB - International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org

ER -

Wand M, Schultz T, Schmidhuber J. Domain-adversarial training for session independent EMG-based speech recognition. In Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH. International Speech Communication Association4 Rue des Fauvettes - Lous TourilsBaixas66390publication@isca-speech.org. 2018. p. 3167-3171 doi: 10.21437/Interspeech.2018-2318

Domain-adversarial training for session independent EMG-based speech recognition

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this