EUVIMEDEUVIMED entdecken
Crossref

Dual-path magnitude-phase learning with bidirectional cross-attention for bone-conducted speech restoration

Dianzhe Ding, Sichen Liu, Junfeng Li iD, Feiran Yang iD

The Journal of the Acoustical Society of America · 01.08.2026 · 160 (2) · 1072-1083

DOI: 10.1121/10.0044570 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of the Acoustical Society of America, online erschienen am 05.08.2026. Verfasst von Dianzhe Ding und 3 weiteren Autor:innen. Beteiligte Einrichtung: Laboratory of Noise and Audio Research, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China. Hauptthemen laut MEDLINE: Bone Conduction, Speech Perception, Speech Acoustics. Die Arbeit nennt 45 Literaturangaben. Gefördert von: Natural Science Foundation of Beijing Municipality, National Natural Science Foundation of China, National Science Fund for Distinguished Young Scholars.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Bone-conducted speech is not subject to background noise but suffers from limited bandwidth. In very low signal-to-noise ratio conditions, it is highly desired to restore air-conducted speech from bone-conducted speech. This paper presents a dual-path frequency-domain U-shaped Convolutional Network (UNet) model to convert bone-conducted speech into high-quality air-conducted speech, which jointly utilizes the magnitude and phase spectrograms of bone-conducted speech. Compared with the widely adopted single-path model for bone-conducted speech restoration and dual-path model for speech denoising, the proposed model consists of dual-path convolution modules in the encoder and decoder and a bidirectional cross-attention module in the bottleneck layer. Dual-path convolution modules utilize two individual branches to extract the magnitude and phase features from the bone-conducted speech spectrogram, which aims to fully exploit the information intrinsic to each spectrogram. Because the two features provide distinct information regarding bone-conducted speech, a bidirectional cross-attention module is introduced to help the model learn consistent information from both features. Moreover, we propose a frequency-weighted spectrogram loss as the training objective. Experiments show that the proposed model outperforms existing bone-conducted speech restoration models, while utilizing only 7.6 M parameters and 2.35 G multiply-accumulate operations, which is lower than that of most competitive methods.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Bone ConductionSpeech PerceptionSpeech AcousticsHumansSignal-To-Noise RatioNoiseSound SpectrographyConvolutional Neural Networks

Autor:innen und Institutionen

  • Laboratory of Noise and Audio Research, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China.
  • University of Chinese Academy of Sciences, Beijing 100049, China.
  • School of Advanced Technology, Xi'an Jiaotong-Liverpool University, Suzhou, Jiangsu, China.
  • Key Laboratory of Speech Acoustics and Content Understanding, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China.
  • State Key Laboratory of Acoustics and Marine Information, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China.

ORCID-Profile: Junfeng Li, Feiran Yang

Förderung

  • Natural Science Foundation of Beijing Municipality
  • National Natural Science Foundation of China
  • National Science Fund for Distinguished Young Scholars

Zitieren

APA

Ding, D., Liu, S., Li, J., & Yang, F. (2026). Dual-path magnitude-phase learning with bidirectional cross-attention for bone-conducted speech restoration. The Journal of the Acoustical Society of America, 160(2), 1072-1083. https://doi.org/10.1121/10.0044570

Vancouver

Ding D, Liu S, Li J, Yang F. Dual-path magnitude-phase learning with bidirectional cross-attention for bone-conducted speech restoration. The Journal of the Acoustical Society of America. 2026;160(2):1072-1083. doi:10.1121/10.0044570

BibTeX
@article{Ding2026,
  author = {Ding, Dianzhe and Liu, Sichen and Li, Junfeng and Yang, Feiran},
  title = {Dual-path magnitude-phase learning with bidirectional cross-attention for bone-conducted speech restoration},
  journal = {The Journal of the Acoustical Society of America},
  year = {2026},
  volume = {160},
  number = {2},
  pages = {1072--1083},
  doi = {10.1121/10.0044570}
}
Kostenlos starten →