Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Proceedings - 2025 IEEE 36th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2025 |
| Herausgeber (Verlag) | Institute of Electrical and Electronics Engineers Inc. |
| Seiten | 172-173 |
| Seitenumfang | 2 |
| ISBN (elektronisch) | 9798331595524 |
| ISBN (Print) | 979-8-3315-9553-1 |
| Publikationsstatus | Veröffentlicht - 28 Juli 2025 |
| Veranstaltung | 36th IEEE International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2025 - Vancouver, Kanada Dauer: 28 Juli 2025 → 30 Juli 2025 |
Publikationsreihe
| Name | Proceedings of the International Conference on Application-Specific Systems, Architectures and Processors |
|---|---|
| ISSN (Print) | 2160-0511 |
| ISSN (elektronisch) | 2160-052X |
Abstract
Many deep neural networks (DNNs) have been applied lately in the field of speech enhancement. One particular subfield, where DNNs have shifted the boundaries of what is considered possible, is noise reduction, where the degrading effects of sounds interfering with speech are minimized. This is especially relevant for hearing impaired listeners, as their ability to understand speech in noisy circumstances is reduced. In contrast to traditional methods, which are known to improve speech quality, DNNs promise to also improve speech intelligibility. Due to the high computational complexity, DNNs have not yet been deployed on a hearing aid processor, constrained by frequencies up to 50 MHz and memory up to 2 MB. In this work we deploy a convolutional neural network (CNN) trained for noise reduction to a hearing-aid system-on-chip (SoC) developed at our institute. Real time capability is achieved by thorough optimization of the C -Code, leading to a speed up by a factor of 88 for the inference relevant layers when compared to a naïve C-Code implementation. The CNN approach is compared to an implementation of a traditional noise reduction method regarding their speech enhancement performance on white and complex noise and their computational cost. While both methods improve the speech quality measured with Perceptual Evaluation of Speech Quality (PESQ), only the CNN achieves a Short-Time Objective Intelligibility (STOI) improvement of 0.077 for complex noise. On the other hand, the CNN has a higher processor utilization of 60.1% compared to 23.5% for the traditional approach. Nonetheless, both methods are real time capable and consume only 3.3 mW for the CNN and 1.78 mW for the traditional approach, respectively.
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Hardware und Architektur
- Informatik (insg.)
- Computernetzwerke und -kommunikation
Ziele für nachhaltige Entwicklung
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Proceedings - 2025 IEEE 36th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2025. Institute of Electrical and Electronics Engineers Inc., 2025. S. 172-173 (Proceedings of the International Conference on Application-Specific Systems, Architectures and Processors).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Noise Reduction in Hearing-Aid Processors
T2 - 36th IEEE International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2025
AU - Klein, Simon
AU - Rossol, Lando
AU - Venema, Finn
AU - Schonewald, Sven
AU - Karrenbauer, Jens
AU - Blume, Holger
N1 - Publisher Copyright: © 2025 IEEE.
PY - 2025/7/28
Y1 - 2025/7/28
N2 - Many deep neural networks (DNNs) have been applied lately in the field of speech enhancement. One particular subfield, where DNNs have shifted the boundaries of what is considered possible, is noise reduction, where the degrading effects of sounds interfering with speech are minimized. This is especially relevant for hearing impaired listeners, as their ability to understand speech in noisy circumstances is reduced. In contrast to traditional methods, which are known to improve speech quality, DNNs promise to also improve speech intelligibility. Due to the high computational complexity, DNNs have not yet been deployed on a hearing aid processor, constrained by frequencies up to 50 MHz and memory up to 2 MB. In this work we deploy a convolutional neural network (CNN) trained for noise reduction to a hearing-aid system-on-chip (SoC) developed at our institute. Real time capability is achieved by thorough optimization of the C -Code, leading to a speed up by a factor of 88 for the inference relevant layers when compared to a naïve C-Code implementation. The CNN approach is compared to an implementation of a traditional noise reduction method regarding their speech enhancement performance on white and complex noise and their computational cost. While both methods improve the speech quality measured with Perceptual Evaluation of Speech Quality (PESQ), only the CNN achieves a Short-Time Objective Intelligibility (STOI) improvement of 0.077 for complex noise. On the other hand, the CNN has a higher processor utilization of 60.1% compared to 23.5% for the traditional approach. Nonetheless, both methods are real time capable and consume only 3.3 mW for the CNN and 1.78 mW for the traditional approach, respectively.
AB - Many deep neural networks (DNNs) have been applied lately in the field of speech enhancement. One particular subfield, where DNNs have shifted the boundaries of what is considered possible, is noise reduction, where the degrading effects of sounds interfering with speech are minimized. This is especially relevant for hearing impaired listeners, as their ability to understand speech in noisy circumstances is reduced. In contrast to traditional methods, which are known to improve speech quality, DNNs promise to also improve speech intelligibility. Due to the high computational complexity, DNNs have not yet been deployed on a hearing aid processor, constrained by frequencies up to 50 MHz and memory up to 2 MB. In this work we deploy a convolutional neural network (CNN) trained for noise reduction to a hearing-aid system-on-chip (SoC) developed at our institute. Real time capability is achieved by thorough optimization of the C -Code, leading to a speed up by a factor of 88 for the inference relevant layers when compared to a naïve C-Code implementation. The CNN approach is compared to an implementation of a traditional noise reduction method regarding their speech enhancement performance on white and complex noise and their computational cost. While both methods improve the speech quality measured with Perceptual Evaluation of Speech Quality (PESQ), only the CNN achieves a Short-Time Objective Intelligibility (STOI) improvement of 0.077 for complex noise. On the other hand, the CNN has a higher processor utilization of 60.1% compared to 23.5% for the traditional approach. Nonetheless, both methods are real time capable and consume only 3.3 mW for the CNN and 1.78 mW for the traditional approach, respectively.
KW - Deep Neural Networks
KW - Hearing Aids
KW - Noise Reduction
KW - SmartHeaP
KW - Speech Enhancement
UR - http://www.scopus.com/inward/record.url?scp=105015844270&partnerID=8YFLogxK
U2 - 10.1109/ASAP65064.2025.00037
DO - 10.1109/ASAP65064.2025.00037
M3 - Conference contribution
AN - SCOPUS:105015844270
SN - 979-8-3315-9553-1
T3 - Proceedings of the International Conference on Application-Specific Systems, Architectures and Processors
SP - 172
EP - 173
BT - Proceedings - 2025 IEEE 36th International Conference on Application-Specific Systems, Architectures and Processors, ASAP 2025
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 28 July 2025 through 30 July 2025
ER -