南方科技大学知识苑(SUSTech KC): 基于智能听诊器的肺炎诊断和感染程度评估

题名	基于智能听诊器的肺炎诊断和感染程度评估
其他题名	PNEUMONIA DIAGNOSIS AND SEVERITY ASSESSMENT USING INTELLIGENT WIRELESS STETHOSCOPE
姓名	吴昌义
姓名拼音	WU Changyi
学号	12232616
学位类型	硕士
学位专业	0856 材料与化工
学科门类/专业学位类别	08 工学
导师	王文锦
导师单位	生物医学工程系
外机构导师	樊正伟
外机构导师单位	易方达资产管理有限公司
论文答辩日期	2024-05-14
论文提交日期	2024-06-13
学位授予单位	南方科技大学
学位授予地点	深圳
摘要	由于长期以来肺部疾病监测机制的不完善，肺炎已经成为全球主要致死原因之一。目前，在新冠疫情常态化的背景下，由疾病引发的肺炎病例数量持续增加，对公共卫生系统构成了重大挑战。在此背景下，提高肺炎的早期诊断和治疗效率尤为重要。传统的肺炎诊断方法需要昂贵的设备和专业人员操作，这对不发达地区的患者而言并不友好。随着人工智能和机器学习技术的发展，自动化肺炎诊断系统正在成为可能。这些系统通过分析患者的医学影像、生理信号和临床数据，能够辅助医生进行更快速、更准确的诊断。在本研究中，我们开展了一项临床试验，旨在探索基于肺部呼吸音的智能听诊器在肺炎诊断中的应用潜力。该试验共纳入 108 例受试者，其中肺炎患者与非肺炎患者各 54 例。我们采集了每位受试者六个肺部关键位置的呼吸音，并获取了相应部位的 CT 图像。在此基础上，我们构建了一个先进的肺部呼吸音识别框架，并对预处理算法进行了优化，以确保捕捉每个呼吸周期的完整信息。此外，我们还引入了全注意力机制，以实现对任意长度上下文信息的高效提取。本研究的另一重要贡献在于量化了不同感染程度肺炎患者的呼吸音特征，并据此开发了一个自动化的肺炎严重程度评估系统。临床验证部分包括两个关键任务：区分肺炎患者与非肺炎患者，以及区分肺炎患者的感染程度，包括健康状态、轻度肺炎和重度肺炎的区分。根据在公开数据集上的预实验结果，我们的预处理方法显著提升了模型对异常呼吸音识别的敏感性。此外，我们引入的全自注意力机制相对其他模型取得最佳性能证实其在肺音识别上的有效性。在最终的基于肺部呼吸音的肺炎诊断和感染程度评估实验结果中，我们发现，综合考虑肺部多个部位呼吸音信息的联合诊断模型，在性能上超越了仅依赖单一部位肺部呼吸音信息的模型。具体而言，我们提出的 AST 模型在综合六个肺部呼吸音信息的设置下，实现了77.61% 的诊断准确率和 79.16% 的召回率。而 VGG16 模型在肺炎严重程度评估任务中表现出色，实现 52.12% 的准确率和 49.75% 的召回率，证明使用肺音来进行肺炎感染程度的可行性。本研究为无线智能听诊器诊断肺炎提供了科学依据，并开辟了实时监测肺炎的新方法，未来有望实现更精准的病情评估和个性化治疗。
其他摘要	Clinical research indicates that pneumonia has become a leading cause of death worldwide due to inadequate surveillance. In the midst of the ongoing COVID-19 pandemic, rising pneumonia cases are challenging public health systems and highlighting the need for effective early diagnosis and treatment. Traditional pneumonia diagnosis requires expensive equipment and specialised staff, making it inaccessible in underdeveloped regions. However, advances in artificial intelligence and machine learning are enabling automated diagnostic systems. These systems help doctors make faster and more accurate diagnoses by analysing medical images, physiological signals and clinical data. In this study, we conducted a clinical trial to evaluate the potential of an intelligent stethoscope for diagnosing pneumonia based on respiratory sounds. We enrolled 108 participants, 54 with pneumonia and 54 without pneumonia. Respiratory sounds were collected from six key lung locations for each participant, along with corresponding CT images. We developed an advanced respiratory sound recognition framework and optimized preprocessing algorithms to ensure complete respiratory cycle capture. Additionally, a full attention mechanism was introduced for efficient context extraction. We quantified respiratory sound characteristics across different pneumonia severities and developed an automated severity assessment system. The clinical validation focused on two tasks: distinguishing between pneumonia and non pneumonia patients and categorizing pneumonia severity into healthy, mild, and severe states. Our study’s results showed that our preprocessing method significantly improved sensitivity to abnormal respiratory sounds on a public dataset. The self-attention mechanism we introduced also proved highly effective in respiratory sound recognition. The final results indicate that the multi-site model, integrating information from various lung positions, outperforms the single-site model in predicting pneumonia severity. Specifically, AST achieved the best performance with 77.61% precision and 79.16% recall in the six-site setting, while VGG16 attained the best performance in disease severity as sessment with 52.12% precision and 49.75% recall. This study provides valuable insights and methods for diagnosing and monitoring lung diseases using wireless stethoscopes.
关键词	肺部呼吸音无线听诊器肺炎感染程度
其他关键词	Respiratory Sounds Wireless Stethoscope Pneumonia Infection Severity
语种	中文
培养类别	独立培养
入学年份	2022
学位授予年份	2024-06
参考文献列表	[1] HUANG D M, HUANG J, QIAO K, et al. Deep learning-based lung sound analysis for intelligent stethoscope[J]. Military Medical Research, 2023, 10: 20-44. [2] HUANG D, WANG L, WANG W. A multi-center clinical trial for wireless stethoscope-based diagnosis and prognosis of children community-acquired pneumonia[J]. IEEE Transactions on Biomedical Engineering, 2023, 70: 2215-2226. [3] FRANCONE M, IAFRATE F, MASCI G M, et al. Chest CT score in COVID-19 patients:Correlation with disease severity and short-term prognosis[J]. European Radiology, 2020, 30: 6808-6817. [4] The Top 10 Causes of Death[EB/OL]. World Health Organization, [2024-03-01]. https://www.who.int/zh/news-room/fact-sheets/detail. [5] KAMEPALLI S, RAO B S, RAO N C S. Custom-built deep convolutional neural networkfor breathing sound classification to detect respiratory diseases[C]//International Conference on Computational Intelligence and Data Engineering (ICCIDE). 2022: 189-201. [6] KUMAR S, BHAGAT V, SAHU P, et al. A novel multimodal framework for early diagnosis and classification of COPD based on CT scan images and multivariate pulmonary respiratory diseases[J]. Computer Methods and Programs in Biomedicine, 2024, 243: 107909-107911. [7] HUANG B, SONG Y, CUI Z, et al. Gravitational search algorithm-extreme learning machine for COVID-19 active cases forecasting[J]. IET Software, 2023, 17: 554-565. [8] SANGLE S, GAIKWAD C. COVID-19 detection using spectral and statistical features of cough and breath sounds[C]//2021 International Conference on Decision Aid Sciences and Application (DASA). 2021: 182-186. [9] ANDRONIKOU S, GOUSSARD P, SORANTIN E. Computed tomography in children withcommunity-acquired pneumonia[J]. Pediatric Radiology, 2017, 47: 1431-1440. [10] HUANG D, WANG L, LU H, et al. A contrastive embedding-based domain adaptation method for lung sound recognition in children community-acquired pneumonia[C]//2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2023: 1-5. [11] SHUVO S B, ALI S N, SWAPNIL S I, et al. A lightweight CNN model for detecting respiratory diseases from lung auscultation sounds using EMD-CWT-based hybrid scalogram [J]. IEEE Journal of Biomedical and Health Informatics, 2021, 25: 2595-2603. [12] KWON A M, KANG K. A temporal dependency feature in lower dimension for lung sound signal classification[J]. Scientific Reports, 2022, 12: 7877-7889. [13] BRUNESE L, MERCALDO F, REGINELLI A, et al. A neural network-based method forrespiratory sound analysis and lung disease detection[J]. Applied Sciences, 2022, 12: 3861-3877. [14] 4.4 million+ Stunning Free Images to Use Anywhere[EB/OL]. Pixabay, [2024-03-01]. https://pixabay.com. [15] SUBASINGHE A, ABEYWICKRAMA A, DISSANAYAKE H, et al. Smart stethoscope: Intelligent respiratory disease prediction system[C]//2022 2nd International Conference on Advanced Research in Computing (ICARC). 2022: 242-247. [16] TARIQ Z, SHAH S K, LEE Y. Feature-based fusion using CNN for lung and heart sound classification[J]. Sensors, 2022, 22: 1492-1521. [17] DROUIN E, CHAMBELLAN A, HAUTECOEUR P. Laennecs’mediate auscultation: Child’s play[J]. American Heart Journal, 2023, 255: 52-57. [18] WEINBERG F. The history of the stethoscope[J]. Canadian Family Physician, 1993, 39: 2223-2224. [19] GROOBY E, SITAULA C, FATTAHI D, et al. Noisy neonatal chest sound separation for high quality heart and lung sounds[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 27: 2635-2646. [20] NGUYEN T, PERNKOPF F. Lung sound classification using co-tuning and stochastic normalization[J]. IEEE Transactions on Biomedical Engineering, 2022, 69: 2872-2882. [21] TRIPATHY R K, DASH S, RATH A, et al. Automated detection of pulmonary diseases from lung sound signals using fixed-boundary-based empirical wavelet transform[J]. IEEE Sensors Letters, 2022, 6: 1-4. [22] DATTA S, DUTTA CHOUDHURY A, DESHPANDE P, et al. Automated lung sound analysis for detecting pulmonary abnormalities[C]//2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2017: 4594-4598. [23] FERNANDO T, SRIDHARAN S, DENMAN S, et al. Robust and interpretable temporal convolution network for event detection in lung sound recordings[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26: 2898-2908. [24] ZHAO Z, GONG Z, NIU M, et al. Automatic respiratory sound classification via multi-branch temporal convolutional network[C]//2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2022: 9102-9106. [25] YU S, DING Y, QIAN K, et al. A glance-and-gaze network for respiratory sound classification[C]//2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2022: 9007-9011. [26] PETMEZAS G, CHEIMARIOTIS G A, STEFANOPOULOS L, et al. Automated lung sound classification using a hybrid CNN-LSTM network and focal loss function[J]. Sensors, 2022, 22: 1218-1232. [27] SONG W, HAN J, SONG H. Contrastive embeddind learning method for respiratory sound classification[C]//2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2021: 1275-1279. [28] HUANG H, YANG D, YANG X, et al. Portable multifunctional electronic stethoscope[C]//2019 IEEE 3rd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC). 2019: 691-694. [29] KLUM M, LEIB F, OBERSCHELP C, et al. Wearable multimodal stethoscope patch for wireless biosignal acquisition and long-term auscultation[C]//2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2019: 5781-5785. [30] MALWADE J, SAYYED S, NASIR J, et al. Wireless stethoscope with bluetooth technology [C]//2020 International Conference on Computational Performance Evaluation (ComPE). 2020: 168-172. [31] JOSHI S S, PATIL M R, KANAWADE N P, et al. Bluetooth-based wireless digital stethoscope [C]//2021 International Conference on Emerging Smart Computing and Informatics (ESCI). 2021: 197-202. [32] SZOT S, LEVIN A, RAGAZZI A, et al. A wireless digital stethoscope design[C]//2018 14th IEEE International Conference on Signal Processing (ICSP). 2018: 74-78. [33] JOSHITHA C, KANAKARAJA P, ROOBAN S, et al. Design and implementation of Wi-Fi enabled contactless electronic stethoscope[C]//2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS). 2022: 932-936. [34] RAMESHA M, DANKANGOWDA V, JEEVAN K, et al. Implementation of IoT based wireless electronic stethoscope[C]//2020 Third International Conference on Multimedia Processing, Communication & Information Technology (MPCIT). 2020: 103-106. [35] HARSHITHA R, PARAMESHACHARI B, YATHIRAJ G, et al. A deep learning model using VGG-16 and neural networks to identify pneumonia from chest X-ray images[C]//2024 International Conference on Integrated Circuits and Communication Systems (ICICACS). 2024: 1-6. [36] RAO D R, PATHAKAMUDI B, DAKSHAYANI R. An approach to avoid SSS problem in the data space using Fisher Face (PCA+ LDA) technique: A case study on chest X-ray pneumonia data[C]//2024 IEEE International Students’ Conference on Electrical, Electronics and Computer Science (SCEECS). 2024: 1-6. [37] AMRULLOH Y A, MAULIDIN L M. Spectral analysis of abnormal breath sounds in childhood pneumonia[C]//2018 International Symposium on Electronics and Smart Devices (ISESD). 2018: 1-5. [38] SOLÀ J, BRAUN F, MUNTANÉ E, et al. Towards an unsupervised device for the diagnosis of childhood pneumonia in low resource settings: Automatic segmentation of respiratory sounds[C]//2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2016: 283-286. [39] AMOSE J, MANIMEGALAI P, NARMATHA C, et al. Comparative performance analysis of kernel functions in support vector machines in the diagnosis of pneumonia using lung sounds [C]//2022 2nd International Conference on Computing and Information Technology (ICCIT). 2022: 320-324. [40] FRAIWAN M, FRAIWAN L, KHASSAWNEH B, et al. A dataset of lung sounds recorded from the chest wall using an electronic stethoscope[J]. Data in Brief, 2021, 35: 106899-106913. [41] GROOBY E, HE J, FATTAHI D, et al. A new non-negative matrix co-factorisation approach for noisy neonatal chest sound separation[C]//2021 43rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2021: 5668-5673. [42] SINGH D, SINGH B K, BEHERA A K. Comparative analysis of lung sound denoising technique[C]//2020 First International Conference on Power, Control and Computing Technologies (ICPC2T). 2020: 406-410. [43] CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[J]. Computer Science, 2014:1062-1078. [44] HARVILL J, WANI Y, ALAM M, et al. Estimation of respiratory rate from breathing audio[C]//2022 44th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2022: 4599-4603. [45] ZHANG Q, FRANCISCO C, KABIR M, et al. Noise removal of tracheal sound recorded during CPET to determine respiratory rate[C]//2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2019: 4650-4653. [46] LIN I, SER W, ZHANG J, et al. A signal-noise separation algorithm for the estimation of respiratory rate from breath sound[C]//2011 8th International Conference on Information, Communications and Signal Processing (ICICS). 2011: 1-4. [47] HSIAO C H, LIN T W, LIN C W, et al. Breathing sound segmentation and detection using transfer learning techniques on an attention-based encoder-decoder architecture[C]//2020 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2020: 754-759. [48] JÁCOME C, RAVN J, HOLSBØ E, et al. Convolutional neural network for breathing phase detection in lung sounds[J]. Sensors, 2019, 19: 1789-1798. [49] MCLANE I, LAUWERS E, STAS T, et al. Comprehensive analysis system for automated respiratory cycle segmentation and crackle peak detection[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26: 1847-1860. [50] PAN F, YE T, SUN P, et al. Time course of lung changes at chest CT during recovery from coronavirus disease 2019 (COVID-19)[J]. Radiology, 2020, 295: 715-721. [51] PRAMONO R X A, BOWYER S, RODRIGUEZ-VILLEGAS E. Automatic adventitious respiratory sound analysis: A systematic review[J]. PloS One, 2017, 12: 177881-177926. [52] NGUYEN T, PERNKOPF F. Lung sound classification using snapshot ensemble of convolutional neural networks[C]//2020 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2020: 760-763. [53] İÇER S, GENGEÇ Ş. Classification and analysis of non-stationary characteristics of crackle and rhonchus lung adventitious sounds[J]. Digital Signal Processing, 2014, 28: 18-27. [54] ROCHA B M, PESSOA D, MARQUES A, et al. Automatic classification of adventitious respiratory sounds: A (un) solved problem?[J]. Sensors, 2020, 21: 37-57. [55] SHARMA G, UMAPATHY K, KRISHNAN S. Trends in audio signal feature extraction methods[J]. Applied Acoustics, 2020, 158: 107020-107040. [56] BROWN C, CHAUHAN J, GRAMMENOS A, et al. Exploring automatic diagnosis of COVID-19 from crowdsourced respiratory sound data[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD). 2020: 3474-3484. [57] XU L, CHENG J, LIU J, et al. Arsc-net: Adventitious respiratory sound classification network using parallel paths with channel-spatial attention[C]//2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). 2021: 1125-1130. [58] GAIROLA S, TOM F, KWATRA N, et al. Respirenet: A deep neural network for accurately detecting abnormal lung sounds in limited data setting[C]//2021 43rd Annual International Con ference of the IEEE Engineering in Medicine and Biology Society (EMBC). 2021: 527-530. [59] ROCHA B M, FILOS D, MENDES L, et al. An open access database for the evaluation of respiratory sound classification algorithms[J]. Physiological Measurement, 2019, 40: 035001-035031. [60] SHI L, ZHANG J, YANG B, et al. Lung sound recognition method based on multi-resolution interleaved net and time-frequency feature enhancement[J]. IEEE Journal of Biomedical and Health Informatics, 2023, 27: 4768-4779. [61] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the IEEE International Conference on Computer Vision (ICCV). 2017: 2980-2988. [62] DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words:Transformers for image recognition at scale[C]//International Conference on Learning Repre sentations (ICLR). 2021: 11929-11952. [63] GONG Y, CHUNG Y A, GLASS J. Ast: Audio spectrogram transformer[C]//2021 Annual Conference of the International Speech Communication Association (Interspeech). 2021: 692-698. [64] REHMER A, KROLL A. On the vanishing and exploding gradient problem in gated recurrent units[J]. IFAC-PapersOnLine, 2020, 53: 1243-1248. [65] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. Advances inNeural Information Processing Systems, 2017, 30: 03750-03762. [66] WU C, HUANG D, TAO X, et al. Intelligent stethoscope using full self-attention mechanism for abnormal respiratory sound recognition[C]//2023 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI). 2023: 1-4. [67] CHEN Z, WANG H, YEH C H, et al. Classify respiratory abnormality in lung sounds using STFT and a fine-tuned ResNet18 network[C]//2022 IEEE Biomedical Circuits and Systems Conference (BioCAS). 2022: 233-237. [68] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//International Conference on Learning Representations (ICLR). 2015: 1556-1570.
所在学位评定分委会	材料与化工
国内图书分类号	TP391.4
来源库	人工提交
成果类型	学位论文
条目标识符	http://sustech.caswiz.com/handle/2SGJ60CL/765614
专题	南方科技大学工学院_生物医学工程系
推荐引用方式 GB/T 7714	吴昌义. 基于智能听诊器的肺炎诊断和感染程度评估[D]. 深圳. 南方科技大学,2024.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可	操作
12232616-吴昌义-生物医学工程系（11448KB）	--	--	限制开放	--	请求全文