WBRCNet 用于呼吸音分类的连续小波变换图像预处理方法
WBRCNet for respiratory sound classification using continuous wavelet scalograms and blank region cropping
研究开发了 WBRCNet 网络,通过优化连续小波时频表示来提高呼吸音周期级分类的准确性。该方法在 ICBHI 2017 数据集上实现了 48.88% 的敏感性和 76.90% 的特异性,ICBHI 得分为 62.89%。尽管取得了进展,但受限于较低的敏感性和缺乏外部验证,目前尚不能用于可靠的临床筛查或部署。
- Zhibin Zhang1,
- Tao Gu ORCID: orcid.org/0009-0004-3470-42571,
- Chenxi Luo2,
- Yan Zhang1,
- Wenjie Wen1 &
- …
- Chao Zhang1
We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.
Abstract
Respiratory sounds are non-stationary biomedical signals, and automatic classification remains challenging because of limited samples, variable cycle lengths and complex abnormal acoustic patterns. This study aimed to improve cycle-level respiratory sound classification by optimizing continuous wavelet time-frequency representations. A Wavelet-based Blank Region Cropping Network (WBRCNet) was developed on the ICBHI 2017 dataset. Respiratory cycles were resampled, length-normalized using Smart Padding, transformed into continuous wavelet scalograms, and processed by Blank Region Cropping for Wavelet (BRCW) to remove low-energy redundant regions. ResNet34 was used as the classification backbone, and comparative experiments were conducted for time-frequency representation, BRCW, boundary artifact suppression and mother wavelet selection. Continuous wavelet scalograms achieved a higher ICBHI Score than Mel spectrograms in the evaluated comparison. With CWT and class-balanced augmentation held constant, replacing conventional BRC with BRCW increased specificity, sensitivity, and the ICBHI Score. The final model achieved Se of 48.88%, Sp of 76.90%, and an ICBHI Score of 62.89%. WBRCNet provides a scalogram preprocessing strategy for respiratory sound classification on ICBHI 2017. However, limited sensitivity and the absence of external validation prevent conclusions about reliable clinical screening or deployment readiness.
Funding
This research was supported by the Fundamental Research Funds for the Central Universities Graduate Science and Technology Innovation Fund (Grant number: ZY20250111) and the University of Emergency Management Project (Grant number: 062204).
Author information
Authors and Affiliations
School of Computer Science and Engineering, North China Institute of Science and Technology, Beijing, 101601, China
Zhibin Zhang, Tao Gu, Yan Zhang, Wenjie Wen & Chao Zhang
Center of Tissue Engineering and Stem Cells, Guizhou Medical University, Guiyang, 561113, China
Chenxi Luo
Authors
- Zhibin Zhang
- Tao Gu
- Chenxi Luo
- Yan Zhang
- Wenjie Wen
- Chao Zhang
Corresponding author
Correspondence to Tao Gu.
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical approval
Ethical approval was not required for this study because it used a publicly available and anonymized respiratory sound dataset.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Zhang, Z., Gu, T., Luo, C. et al. WBRCNet for respiratory sound classification using continuous wavelet scalograms and blank region cropping. Sci Rep (2026). https://doi.org/10.1038/s41598-026-76165-w
Received: 22 July 2026
Accepted: 12 October 2026
Published: 19 November 2026
DOI: https://doi.org/10.1038/s41598-026-76165-w
Keywords
来源:Nature 健康护理主题 · nature.com