端到端KAN卷积在房颤心音识别中的应用

doi:10.11896/jsjkx.250100102

计算机科学 ›› 2025, Vol. 52 ›› Issue (9): 62-70.doi: 10.11896/jsjkx.250100102

端到端KAN卷积在房颤心音识别中的应用

邓洪¹, 陈燕², 杨宏波³, 赵峰², 蒋永卓¹, 郭涛³, 王威廉¹

1 云南大学信息学院昆明 650504
2 中国人民解放军总医院第一医学中心北京 100853
3 云南省阜外心血管医院昆明 650102

收稿日期:2025-01-15 修回日期:2025-04-13 出版日期:2025-09-15 发布日期:2025-09-11
通讯作者: 王威廉(wlwang_47@126.com)
作者简介:(dh2904799190@163.com)
基金资助:
国家自然科学基金(82172185,81960067);2018年云南省重大科技专项(2018ZF017)

Application of End-to-End Convolutional Kolmogorov-Arnold Networks in Atrial Fibrillation Heart Sound Recognition

DENG Hong¹, CHEN Yan², YANG Hongbo³, ZHAO Feng², JIANG Yongzhuo¹, GUO Tao³, WANG Weilian¹

1 School of Information Science and Engineering,Yunnan University,Kunming 650504,China
2 The First Medical Center of Chinese PLA General Hospital,Beijing 100853,China
3 Fuwai Yunnan Hospital,Chinese Academy of Medical Sciences,Kunming 650102,China

Received:2025-01-15 Revised:2025-04-13 Online:2025-09-15 Published:2025-09-11
About author:DENG Hong,born in 2000,postgra-duate.His main research interests include biomedical signal processing and deep learning.
WANG Weilian,born in 1947,professor.His main research interests include biological signal processing,digital-analog hybrid IC and ASIC design.
Supported by:
National Natural Science Foundation of China(82172185,81960067) and Major Science and Technology Projects of Yunnan Province in 2018(2018ZF017).

摘要/Abstract

摘要： 房颤作为一种严重的心律失常疾病,及早诊断至关重要。房颤的传统检查方法是由心脏科医生借助心电图、超声心动图等设备做出诊断结论。为了缓解传统诊断方法检查成本高、过多依赖临床经验和便捷性不足等问题,创新性地应用Kolmogo-rov-Arnold Network(KAN)来构建房颤心音分析模型。文中探索了KAN卷积在房颤心音识别中的应用,引入了具有灵活线性激活函数和优异参数效率的KAN卷积架构,提出了一种基于KAN卷积的端到端房颤心音识别模型。为提高信号的可用性,首先对心音信号进行预处理,包括心音分割、心音信号的质量评估和数据清洗;然后利用KAN的卷积层、池化层等自动学习;最后采用KAN卷积分类器进行识别研究。在特征提取阶段引入了KAN卷积的自注意力机制和焦点调制,以高效提取信号特征;在分类器阶段研究了KAN卷积的瓶颈结构和正则化手段,以提升模型的识别能力。该模型在中国人民解放军总医院第一医学中心的心音信号数据集上进行了正常和房颤的识别测试,准确率为97.86%,灵敏度为98.18%,特异度为97.43%,F_β值为98.06%。实验结果表明,KAN卷积模型在辅助诊断房颤信号上具有显著的优势。

关键词: 房颤, KAN卷积, 心音识别, 端到端, 数据清洗

Abstract: Atrial fibrillation(AF) is a severe cardiac arrhythmia that requires early diagnosis.Traditional diagnostic methods typically involve cardiologists using electrocardiograms(ECG) and echocardiograms to make diagnostic conclusions.To address issues such as high costs,excessive reliance on clinical expertise,and limited accessibility,this study proposes an innovative application of Kolmogorov-Arnold Networks(KAN) in AF heart sound analysis.This study explores the application of convolutional KAN(CKAN) in AF heart sound recognition,proposing an end-to-end AF identification framework based on CKAN architecture,which incorporates flexible linear activation functions and exhibits superior parameter efficiency.To enhance the usability of heart sound signals,the methodology first applies preprocessing,including signal segmentation,quality assessment,and data cleansing.Subsequently,the model autonomously learns discriminative features through KAN-based convolutional and pooling layers.Fina-lly,a CKAN-based classifier is employed for classification.During the feature extraction phase,self-attention mechanisms and focus modulation are incorporated into CKAN to efficiently extract signal features.In the classification phase,CKAN’s bottleneck structure and regularization techniques are explored to improve the model’s recognition performance.The proposed model is evaluated on a heart sound dataset from The First Medical Center of Chinese PLA General Hospital,achieving an accuracy of 97.86%,sensitivity of 98.18%,specificity of 97.43%,and an F_β score of 98.06%.The results indicate that the CKAN model provides significant advantages in aiding the diagnosis of AF from heart sound signals.

Key words: Atrial fibrillation, Convolutional KAN, Heart sound recognition, End-to-End, Data cleansing

中图分类号:

TN912

邓洪, 陈燕, 杨宏波, 赵峰, 蒋永卓, 郭涛, 王威廉. 端到端KAN卷积在房颤心音识别中的应用[J]. 计算机科学, 2025, 52(9): 62-70. https://doi.org/10.11896/jsjkx.250100102

DENG Hong, CHEN Yan, YANG Hongbo, ZHAO Feng, JIANG Yongzhuo, GUO Tao, WANG Weilian. Application of End-to-End Convolutional Kolmogorov-Arnold Networks in Atrial Fibrillation Heart Sound Recognition[J]. Computer Science, 2025, 52(9): 62-70. https://doi.org/10.11896/jsjkx.250100102

参考文献

[1]GAO M Y,HE L,DU X,et al.20 years of atrial fibrillation epidemiology in China[J].Chinese Journal of Cardiovascular Di-sease,2024,52(2):220-226.
[2]KORNEJ J,BÖRSCHEL C S,BENJAMIN E J,et al.Epidemiology of atrial fibrillation in the 21st century:novel methods and new insights[J].Circulation Research,2020,127(1):4-20.
[3]HAGIWARA Y,FUJITA H,OH S L,et al.Computer-aided diagnosis of atrial fibrillation based on ECG Signals:A review[J].Information Sciences,2018,467:99-114.
[4]TAE-SEOK K,HO-JOONG Y.Role of echocardiography in at-rial fibrillation[J].Journal of Cardiovascular Ultrasound,2011,19(2):51-61.
[5]GIBSON C M,CIAGLO L N,SOUTHARD M C,et al.Diagnos-tic and prognostic value of ambulatory ECG(Holter) monitoring in patients with coronary heart disease:a review[J].Journal of Thrombosis and Thrombolysis,2007,23:135-145.
[6]NABIH-ALI M,EL-DAHSHAN E L S A,YAHIA A S.A review of intelligent systems for heart sound signal analysis[J].Journal of Medical Engineering & Technology,2017,41(7):553-563.
[7]YANG L,LI S,ZHANG Z,et al.Classification of phonocardiogram signals based on envelope optimization model and support vector machine[J].Journal of Mechanics in Medicine and Biology,2020,20(1):1-7.
[8]ZHAO Q H,GENG S J,WANG B Y,et al.Deep Learning for Heart Sound Analysis:A Literature Review[J/OL].https://www.medrxiv.org/content/10.1101/2023.09.16.23295653v1.full-text.
[9]RENNA F,OLIVEIRA J H,COIMBRA M T.Deep Convolutional Neural Networks for Heart Sound Segmentation[J].IEEE Journal of Biomedical and Health Informatics,2019,23(6):2435-2445.
[10]AKRAM,MUHAMMAD USMAN,et al.Analysis of PCG signals using quality assessment and homomorphic filters for localization and classification of heart sounds[J] Computer Methods and Programs in Biomedicine,2018,164:143-157.
[11]WANG Y L,SUN J,YANG H B,et al.Classification model of pulmonary hypertension heart sounds based on time-frequency fusion features[J].Computer Science,2024,51(S1):387-393.
[12]DENG M,MENG T,CAO J,et al.Heart sound classificationbased on improved MFCC features and convolutional recurrent neural networks[J].Neural Networks,2020,130:22-32.
[13]WHANG S E,TAE K H,ROH Y,et al.Responsible AI Challenges in End-to-end Machine Learning[J].IEEE Date Engineering Bulletin,2021,44:79-91.
[14]CHEN Y,SUN Y,LV J,et al.End-to-end heart sound segmentation using deep convolutional recurrent network[J].Complex &Intelligent Systems,2021,7:2103-2117.
[15] LI S,SUN J,YANG H,et al.Interpretable End-to-End heart sound classification[J].Measurement,2024:237.
[16]NAHAR K,AL-HAZAIMEH O M,ABU-EIN A,et al.Phonocardiogram classification based on machine learning withmultiple sound features[J].Journal of Computer Science,2020,16(11):1648-1656.
[17]LIU Z,WANG Y,VAIDYA S,et al.KAN:Kolmogorov-Arnold Networks[J].arXiv:2404.19756,2024.
[18]DROKIN I.Kolmogorov-Arnold Convolutions:Design Principles and Empirical Studies[J].arXiv:2407.01092,2024.
[19]OUYANG D,HE B,GHORBANI A,et al.Video-based AI for beat-to-beat assessment of cardiac function[J].Nature,2020,580(7802):252-256.
[20]ZHAN J,WU X,FU X,et al.Non-contact assessment of cardiac physiology using FO-MVSS-based ballistocardiography:a promising approach for heart failure evaluation[J/OL].https://www.nature.com/articles/s41598-024-53464-8.pdf.
[21]ASMARE M H,WOLDEHANNA F,JANSSENS L,et al.Rheumatic heart disease detection using deep learning from spectro-temporal representation of un-segmented heart sounds[C]//2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society(EMBC).IEEE,2020:168-171.
[22]SINGH S A,MAJUMDER S.Short unsegmented PCG classification based on ensemble classifier[J].Turkish Journal of Electrical Engineering and Computer Sciences,2020,28(2):875-889.
[23]LI F,TANG H,SHANG S,et al.Classification of Heart Sounds Using Convolutional Neural Network[J].Applied Sciences,2020,10(11):3956.
[24]SKALICKY D,KOUCKYU,HADRABA D,et al.Detection of Respiratory Phases in a Breath Sound and Their Subsequent Utilization in a Diagnosis[J].Applied Sciences,2021,11(14):6353.
[25]SATHESH K,KOWSALYA P,ARAVINDRAJ E,et al.Clearer HeartBeats:Enhancement of cardiac sounds using Adaptive Filtering and Wavelet Decomposition[C]//2024 10th International Conference on Communication and Signal Processing(ICCSP).IEEE,2024:947-952.
[26]TANG H,WANG M,HU Y,et al.Automated Signal Quality Assessment for Heart Sound Signal by Novel Features and Evaluation in Open Public Datasets[J].BioMed Research International,2021:7565398.
[27]BODNER A D,TEPSICH A S,SPOLSKI J N,et al.Convolutional Kolmogorov-Arnold Networks[J].arXiv:2406.13155,2024.
[28]ZOU X,XU S,CHEN X,et al.Breaking the von Neumann bottleneck:architecture-level processing-in-memory technology[J].Science China Information Sciences,2021,64(6):1-10.
[29]SRINIVAS A,LIN T Y,PARMAR N,et al.Bottleneck transformers for visual recognition[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2021:16519-16529.
[30]SINGH S,RUWASE O,AWAN A A,et al.A hybrid tensor-expert-data parallelism approach to optimize mixture-of-experts training[C]//Proceedings of the 37th International Conference on Supercomputing.2023:203-214.
[31]LI B,SHEN Y,YANG J,et al.Sparse Mixture-of-Experts are Domain Generalizable Learners[C]//International Conference on Learning Representations(ICLR).2023.
[32]LIU C F,SUN H,DONG H.A study of molecular amplification timing prediction combining Transformer and Kolmogorov Arnold network[J].Journal of Graphics,2024,45(6):1256-1265.
[33]WU H,XIAO B,CODELLA N,et al.Cvt:Introducing convolutions to vision transformers[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2021:22-31.
[34]YANG J,LI C,DAI X,et al.Focal modulation networks[J].Ad-vances in Neural Information Processing Systems,2022,35:4203-4217.
[35]TIAN M D,PENG D T,ZHANG X.Research on sparse optimization with linear inequality constraints based on Huber loss and Capped L1 regularization[J].Theory of Mathematics,2022,12(11):12.
[36]WANG Q,YANG H B,PAN J H,et al.Heart sound classification algorithm based on time-frequency combination feature and adaptive fuzzy neural network[J].Journal of Biomedical Engineering,2023,40(6):1152-1159.
[37]WANG Y L.SUN J.YANG H B,et al.Heart sound classification based on improved mel frequency cepstrum coefficient and integrated decision network method[J]Journal of Blomedical Engineering,2022,39(6):1140-1148.
[38]LOSHCHILOV I,HUTTER F.Decoupled Weight Decay Regularization[J].arXiv.1711.05101,2017.
[39]KIM H,KO J G.Fast Monte-Carlo approximation of the atten-tion mechanism[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:7185-7193.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

端到端KAN卷积在房颤心音识别中的应用

Application of End-to-End Convolutional Kolmogorov-Arnold Networks in Atrial Fibrillation Heart Sound Recognition

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

Metrics

本文评价

推荐阅读 0