SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES

Yu-Chen Lin; Tsun-An Hsieh; Kuo-Hsuan Hung; Cheng Yu; Harinath Garudadri; Yu Tsao; Tei-Wei Kuo

doi:10.1109/ICASSP43922.2022.9747219

SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES

Yu-Chen Lin, Tsun-An Hsieh, Kuo-Hsuan Hung, Cheng Yu, Harinath Garudadri, Yu Tsao, Tei-Wei Kuo

Department of Computer Science

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

4 Citations (Scopus)

Abstract

The incompleteness of speech inputs severely degrades the performance of all the related speech signal processing applications. Although many researches have been proposed to address this issue, they controlled the data missing conditions by simulation with self-defined masking lengths or sizes. Besides, the masking definitions are different among all these experimental settings. This paper presents a novel intermittent speech recovery (ISR) system for real-world self-powered intermittent devices. Three contributive stages: interpolation, enhancement, and combination are applied to the ISR system for speech reconstruction. The experimental results show that our recovery system increases speech quality by up to 591.7%, while increasing speech intelligibility by up to 80.5%. Most importantly, the proposed ISR system improves the WER scores by up to 52.6%. The promising results not only confirm the effectiveness of the reconstruction but also encourage the utilization of these battery-free wearable/IoT devices.

Original language	English
Title of host publication	2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings
Publisher	IEEE
Pages	26-30
ISBN (Electronic)	9781665405409
ISBN (Print)	9781665405416
DOIs	https://doi.org/10.1109/ICASSP43922.2022.9747219
Publication status	Published - 2022
Event	47th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2022) - Hybrid, Marina Bay Sands Expo & Convention Center, Singapore Duration: 22 May 2022 → 27 May 2022 https://2022.ieeeicassp.org/

Publication series

Name	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume	2022-May
ISSN (Print)	1520-6149
ISSN (Electronic)	2379-190X

Conference

Conference	47th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2022)
Abbreviated title	IEEE ICASSP 2022
Place	Singapore
Period	22/05/22 → 27/05/22
Internet address	https://2022.ieeeicassp.org/

Research Keywords

energy harvesting
intermittent systems
internet-of-things
speech recovery
speech signal processing

Access Output

10.1109/ICASSP43922.2022.9747219

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Lin, Y.-C., Hsieh, T.-A., Hung, K.-H., Yu, C., Garudadri, H., Tsao, Y., & Kuo, T.-W. (2022). SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES. In 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings (pp. 26-30). (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2022-May). IEEE. https://doi.org/10.1109/ICASSP43922.2022.9747219

@inproceedings{22932ea7dfa643d88f6d0583217932d1,

title = "SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES",

abstract = "The incompleteness of speech inputs severely degrades the performance of all the related speech signal processing applications. Although many researches have been proposed to address this issue, they controlled the data missing conditions by simulation with self-defined masking lengths or sizes. Besides, the masking definitions are different among all these experimental settings. This paper presents a novel intermittent speech recovery (ISR) system for real-world self-powered intermittent devices. Three contributive stages: interpolation, enhancement, and combination are applied to the ISR system for speech reconstruction. The experimental results show that our recovery system increases speech quality by up to 591.7%, while increasing speech intelligibility by up to 80.5%. Most importantly, the proposed ISR system improves the WER scores by up to 52.6%. The promising results not only confirm the effectiveness of the reconstruction but also encourage the utilization of these battery-free wearable/IoT devices.",

keywords = "energy harvesting, intermittent systems, internet-of-things, speech recovery, speech signal processing",

author = "Yu-Chen Lin and Tsun-An Hsieh and Kuo-Hsuan Hung and Cheng Yu and Harinath Garudadri and Yu Tsao and Tei-Wei Kuo",

year = "2022",

doi = "10.1109/ICASSP43922.2022.9747219",

language = "English",

isbn = "9781665405416",

series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

publisher = "IEEE",

pages = "26--30",

booktitle = "2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings",

address = "United States",

note = "47th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2022), IEEE ICASSP 2022 ; Conference date: 22-05-2022 Through 27-05-2022",

url = "https://2022.ieeeicassp.org/",

}

Lin, Y-C, Hsieh, T-A, Hung, K-H, Yu, C, Garudadri, H, Tsao, Y & Kuo, T-W 2022, SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES. in 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, vol. 2022-May, IEEE, pp. 26-30, 47th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2022), Singapore, 22/05/22. https://doi.org/10.1109/ICASSP43922.2022.9747219

SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES. / Lin, Yu-Chen; Hsieh, Tsun-An; Hung, Kuo-Hsuan et al.
2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings. IEEE, 2022. p. 26-30 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2022-May).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES

AU - Lin, Yu-Chen

AU - Hsieh, Tsun-An

AU - Hung, Kuo-Hsuan

AU - Yu, Cheng

AU - Garudadri, Harinath

AU - Tsao, Yu

AU - Kuo, Tei-Wei

PY - 2022

Y1 - 2022

N2 - The incompleteness of speech inputs severely degrades the performance of all the related speech signal processing applications. Although many researches have been proposed to address this issue, they controlled the data missing conditions by simulation with self-defined masking lengths or sizes. Besides, the masking definitions are different among all these experimental settings. This paper presents a novel intermittent speech recovery (ISR) system for real-world self-powered intermittent devices. Three contributive stages: interpolation, enhancement, and combination are applied to the ISR system for speech reconstruction. The experimental results show that our recovery system increases speech quality by up to 591.7%, while increasing speech intelligibility by up to 80.5%. Most importantly, the proposed ISR system improves the WER scores by up to 52.6%. The promising results not only confirm the effectiveness of the reconstruction but also encourage the utilization of these battery-free wearable/IoT devices.

AB - The incompleteness of speech inputs severely degrades the performance of all the related speech signal processing applications. Although many researches have been proposed to address this issue, they controlled the data missing conditions by simulation with self-defined masking lengths or sizes. Besides, the masking definitions are different among all these experimental settings. This paper presents a novel intermittent speech recovery (ISR) system for real-world self-powered intermittent devices. Three contributive stages: interpolation, enhancement, and combination are applied to the ISR system for speech reconstruction. The experimental results show that our recovery system increases speech quality by up to 591.7%, while increasing speech intelligibility by up to 80.5%. Most importantly, the proposed ISR system improves the WER scores by up to 52.6%. The promising results not only confirm the effectiveness of the reconstruction but also encourage the utilization of these battery-free wearable/IoT devices.

KW - energy harvesting

KW - intermittent systems

KW - internet-of-things

KW - speech recovery

KW - speech signal processing

UR - http://www.scopus.com/inward/record.url?scp=85131239229&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85131239229&origin=recordpage

U2 - 10.1109/ICASSP43922.2022.9747219

DO - 10.1109/ICASSP43922.2022.9747219

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 9781665405416

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 26

EP - 30

BT - 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings

PB - IEEE

T2 - 47th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2022)

Y2 - 22 May 2022 through 27 May 2022

ER -

Lin YC, Hsieh TA, Hung KH, Yu C, Garudadri H, Tsao Y et al. SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES. In 2022 IEEE International Conference on Acoustics, Speech, and Signal Processing - Proceedings. IEEE. 2022. p. 26-30. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). Epub 2022 Apr 27. doi: 10.1109/ICASSP43922.2022.9747219

SPEECH RECOVERY FOR REAL-WORLD SELF-POWERED INTERMITTENT DEVICES

Abstract

Publication series

Conference

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this