{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T21:58:43Z","timestamp":1792792723033,"version":"4.5.0"},"reference-count":33,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2022,8,29]],"date-time":"2022-08-29T00:00:00Z","timestamp":1661731200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2022,8,29]],"date-time":"2022-08-29T00:00:00Z","timestamp":1661731200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["BMC Med Inform Decis Mak"],"abstract":"<jats:title>Abstract<\/jats:title><jats:sec>\n                <jats:title>Background<\/jats:title>\n                <jats:p>The application of machine learning to cardiac auscultation has the potential to improve the accuracy and efficiency of both routine and point-of-care screenings. The use of convolutional neural networks (CNN) on heart sound spectrograms in particular has defined state-of-the-art performance. However, the relative paucity of patient data remains a significant barrier to creating models that can adapt to a wide range of potential variability. To that end, we examined a CNN model\u2019s performance on automated heart sound classification, before and after various forms of data augmentation, and aimed to identify the most optimal augmentation methods for cardiac spectrogram analysis.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Results<\/jats:title>\n                <jats:p>We built a standard CNN model to classify cardiac sound recordings as either normal or abnormal. The baseline control model achieved a PR AUC of 0.763\u2009\u00b1\u20090.047. Among the single data augmentation techniques explored, horizontal flipping of the spectrogram image improved the model performance the most, with a PR AUC of 0.819\u2009\u00b1\u20090.044. Principal component analysis color augmentation (PCA) and perturbations of saturation-value (SV) of the hue-saturation-value (HSV) color scale achieved a PR AUC of 0.779\u2009\u00b1\u2009045 and 0.784\u2009\u00b1\u20090.037, respectively. Time and frequency masking resulted in a PR AUC of 0.772\u2009\u00b1\u20090.050. Pitch shifting, time stretching and compressing, noise injection, vertical flipping, and applying random color filters negatively impacted model performance. Concatenating the best performing data augmentation technique (horizontal flip) with PCA and SV perturbations improved model performance.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Conclusion<\/jats:title>\n                <jats:p>Data augmentation can improve classification accuracy by expanding and diversifying the dataset, which protects against overfitting to random variance. However, data augmentation is necessarily domain specific. For example, methods like noise injection have found success in other areas of automated sound classification, but in the context of cardiac sound analysis, noise injection can mimic the presence of murmurs and worsen model performance. Thus, care should be taken to ensure clinically appropriate forms of data augmentation to avoid negatively impacting model performance.<\/jats:p>\n              <\/jats:sec>","DOI":"10.1186\/s12911-022-01942-2","type":"journal-article","created":{"date-parts":[[2022,8,29]],"date-time":"2022-08-29T08:03:25Z","timestamp":1661760205000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":43,"title":["On the analysis of data augmentation methods for spectral imaged based heart sound classification using convolutional neural networks"],"prefix":"10.1186","volume":"22","author":[{"given":"George","family":"Zhou","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yunchan","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Candace","family":"Chien","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2022,8,29]]},"reference":[{"key":"1942_CR1","doi-asserted-by":"publisher","DOI":"10.1080\/01421590802572767","author":"T Nielsen","year":"2009","unstructured":"Nielsen T, Molgaard H, Ringsted C, Eika B. The development of a new cardiac auscultation test: how do screening and diagnostic skills differ? Med Teach. 2009. https:\/\/doi.org\/10.1080\/01421590802572767.","journal-title":"Med Teach"},{"key":"1942_CR2","doi-asserted-by":"publisher","first-page":"210","DOI":"10.1016\/S0002-9343(00)00673-2","volume":"110","author":"S Mangione","year":"2001","unstructured":"Mangione S. Cardiac auscultatory skills of physicians-in-training: a comparison of three English-speaking countries. Am J Med. 2001;110:210\u20136.","journal-title":"Am J Med"},{"key":"1942_CR3","doi-asserted-by":"publisher","first-page":"236","DOI":"10.1177\/0009922806290028","volume":"46","author":"S Dhuper","year":"2007","unstructured":"Dhuper S, Vashist S, Shah N, Sokal M. Improvement of cardiac auscultation skills in pediatric residents with training. Clin Pediatr (Phila). 2007;46:236\u201340.","journal-title":"Clin Pediatr (Phila)"},{"issue":"6","key":"1942_CR4","doi-asserted-by":"publisher","first-page":"168","DOI":"10.1007\/s10916-019-1286-5","volume":"43","author":"DM Nogueira","year":"2019","unstructured":"Nogueira DM, Ferreira CA, Gomes EF, Jorge AM. Classifying heart sounds using images of motifs, MFCC and temporal features. J Med Syst. 2019;43(6):168. https:\/\/doi.org\/10.1007\/s10916-019-1286-5.","journal-title":"J Med Syst"},{"key":"1942_CR5","doi-asserted-by":"publisher","first-page":"351","DOI":"10.1016\/j.bspc.2017.08.002","volume":"39","author":"M Hamidi","year":"2018","unstructured":"Hamidi M, Ghassemian H, Imani M. Classification of heart sound signal using curve fitting and fractal dimension. Biomed Signal Process Control. 2018;39:351\u20139. https:\/\/doi.org\/10.1016\/j.bspc.2017.08.002.","journal-title":"Biomed Signal Process Control"},{"issue":"8","key":"1942_CR6","doi-asserted-by":"publisher","first-page":"1671","DOI":"10.1088\/1361-6579\/aa7841","volume":"38","author":"V Maknickas","year":"2017","unstructured":"Maknickas V, Maknickas A. Recognition of normal\u2013abnormal phonocardiographic signals using deep convolutional neural networks and mel-frequency spectral coefficients. Physiol Meas. 2017;38(8):1671\u201384.","journal-title":"Physiol Meas"},{"key":"1942_CR7","doi-asserted-by":"publisher","first-page":"012202","DOI":"10.1088\/1742-6596\/953\/1\/012202","volume":"953","author":"D Juniati","year":"2018","unstructured":"Juniati D, Khotimah C, Wardani DEK, Budayasa K. Fractal dimension to classify the heart sound recordings with KNN and fuzzy c-mean clustering methods. J Phys Conf Ser. 2018;953:012202. https:\/\/doi.org\/10.1088\/1742-6596\/953\/1\/012202.","journal-title":"J Phys Conf Ser"},{"issue":"4","key":"1942_CR8","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1007\/s10916-017-0704-9","volume":"41","author":"ME Karar","year":"2017","unstructured":"Karar ME, El-Khafif SH, El-Brawany MA. Automated diagnosis of heart sounds using rule-based classification tree. J Med Syst. 2017;41(4):60. https:\/\/doi.org\/10.1007\/s10916-017-0704-9.","journal-title":"J Med Syst"},{"key":"1942_CR9","doi-asserted-by":"publisher","unstructured":"Deng SW, Han JQ. Towards heart sound classification without segmentation via autocorrelation feature and diffusion maps. Fut Gener Comput Syst. 2016;60:13\u201321. https:\/\/doi.org\/10.1016\/j.future.2016.01.010.","DOI":"10.1016\/j.future.2016.01.010"},{"key":"1942_CR10","doi-asserted-by":"publisher","first-page":"220","DOI":"10.1016\/j.eswa.2017.05.014","volume":"84","author":"W Zhang","year":"2017","unstructured":"Zhang W, Han J, Deng S. Heart sound classification based on scaled spectrogram and tensor decomposition. Expert Syst Appl. 2017;84:220\u201331. https:\/\/doi.org\/10.1016\/j.eswa.2017.05.014.","journal-title":"Expert Syst Appl"},{"issue":"8","key":"1942_CR11","doi-asserted-by":"publisher","first-page":"1701","DOI":"10.1088\/1361-6579\/aa7623","volume":"38","author":"BM Whitaker","year":"2017","unstructured":"Whitaker BM, Suresha PB, Liu C, Clifford GD, Anderson DV. Combining sparse coding and time-domain features for heart sound classification. Physiol Meas. 2017;38(8):1701\u201313. https:\/\/doi.org\/10.1088\/1361-6579\/aa7623.","journal-title":"Physiol Meas"},{"key":"1942_CR12","doi-asserted-by":"crossref","unstructured":"Cheng X, Zhan Q, Wang J, Ma R. A high recognition rate of feature extraction algorithm without segmentation. In: IEEE 6th international conference on industrial engineering and applications (ICIEA); Tokyo: IEEE; 2019. p. 923\u20137.","DOI":"10.1109\/IEA.2019.8714943"},{"issue":"3","key":"1942_CR13","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1007\/s10439-006-9232-3","volume":"35","author":"P Wang","year":"2007","unstructured":"Wang P, Lim CS, Chauhan S, Foo JY, Anantharaman V. Phonocardiographic signal analysis method using a modified hidden Markov model. Ann Biomed Eng. 2007;35(3):367\u201374. https:\/\/doi.org\/10.1007\/s10439-006-9232-3.","journal-title":"Ann Biomed Eng"},{"issue":"2","key":"1942_CR14","doi-asserted-by":"publisher","first-page":"221","DOI":"10.1016\/j.compbiomed.2007.10.006","volume":"38","author":"S Chauhan","year":"2008","unstructured":"Chauhan S, Wang P, Sing Lim C, Anantharaman V. A computer-aided MFCC-based HMM system for automatic auscultation. Comput Biol Med. 2008;38(2):221\u201333. https:\/\/doi.org\/10.1016\/j.compbiomed.2007.10.006.","journal-title":"Comput Biol Med"},{"key":"1942_CR15","doi-asserted-by":"crossref","unstructured":"Yang TI, Hsieh H. Classification of acoustic physiological signals based on deep learning neural networks with augmented features. In: 2016 Computing in cardiology conference (CinC), Vancouver, BC; 2016. p. 569\u201372.","DOI":"10.22489\/CinC.2016.163-228"},{"issue":"21","key":"1942_CR16","doi-asserted-by":"publisher","first-page":"4819","DOI":"10.3390\/s19214819","volume":"19","author":"A Raza","year":"2019","unstructured":"Raza A, Mehmood A, Ullah S, Ahmad M, Choi GS, On BW. Heartbeat sound signal classification using deep learning. Sensors (Basel, Switzerland). 2019;19(21):4819. https:\/\/doi.org\/10.3390\/s19214819.","journal-title":"Sensors (Basel, Switzerland)"},{"key":"1942_CR17","doi-asserted-by":"crossref","unstructured":"Ryu H, Park J, Shin H, Classification of heart sound recordings using convolution neural network. In 2016 Computing in cardiology conference (CinC), Vancouver, BC; 2016. p. 1153\u20136.","DOI":"10.22489\/CinC.2016.329-134"},{"issue":"2017","key":"1942_CR18","doi-asserted-by":"publisher","first-page":"132","DOI":"10.1016\/j.compbiomed.2018.06.026","volume":"100","author":"B Bozkurt","year":"2018","unstructured":"Bozkurt B, Germanakis I, Stylianou Y. A study of time-frequency features for CNN-based automatic heart sound classification for pathology detection. Comput Biol Med. 2018;100(2017):132\u201343.","journal-title":"Comput Biol Med"},{"key":"1942_CR19","doi-asserted-by":"publisher","first-page":"8","DOI":"10.1109\/MIS.2009.36","volume":"24","author":"A Halevy","year":"2009","unstructured":"Halevy A, Norvig P, Pereira F. The unreasonable effectiveness of data. IEEE Intell Syst. 2009;24:8\u201312.","journal-title":"IEEE Intell Syst"},{"key":"1942_CR20","unstructured":"Chen S, Abhinav S, Saurabh S, Abhinav G. Revisiting unreasonable effectiveness of data in deep learning era. In: ICCV; 2017. p. 843\u201352."},{"key":"1942_CR21","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","volume":"6","author":"C Shorten","year":"2019","unstructured":"Shorten C, Khoshgoftaar TM. A survey on image data augmentation for deep learning. J Big Data. 2019;6:60. https:\/\/doi.org\/10.1186\/s40537-019-0197-0.","journal-title":"J Big Data"},{"issue":"9","key":"1942_CR22","doi-asserted-by":"publisher","first-page":"2181","DOI":"10.1088\/0967-3334\/37\/12\/2181","volume":"37","author":"C Liu","year":"2016","unstructured":"Liu C, Springer D, Li Q, Moody B, Juan RA, Chorro FJ, Castells F, Roig JM, Silva I, Johnson AE, Sye Z, Schmidt SE, Papadaniil CD, Hadjileontiadis L, Naseri H, Moukadem A, Dieterlen A, Brandt C, Tang H, Samieinasab M, Samieinasab MR, Sameni R, Mark RG, Clifford GD. An open access database for the evaluation of heart sound algorithms. Physiol Meas. 2016;37(9):2181\u2013213.","journal-title":"Physiol Meas"},{"issue":"23","key":"1942_CR23","doi-asserted-by":"publisher","first-page":"e215","DOI":"10.1161\/01.CIR.101.23.e215","volume":"101","author":"A Goldberger","year":"2000","unstructured":"Goldberger A, Amaral L, Glass L, Hausdorff J, Ivanov PC, Mark R, Mietus JE, Moody GB, Peng CK, Stanley HE. PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals. Circulation. 2000;101(23):e215\u201320.","journal-title":"Circulation"},{"issue":"6","key":"1942_CR24","doi-asserted-by":"publisher","first-page":"1326","DOI":"10.1109\/TBME.2016.2603787","volume":"64","author":"S Kang","year":"2017","unstructured":"Kang S, Doroshow R, McConnaughey J, Shekhar R. Automated identification of innocent Still\u2019s Murmur in children. IEEE Trans Biomed Eng. 2017;64(6):1326\u201334. https:\/\/doi.org\/10.1109\/TBME.2016.2603787.","journal-title":"IEEE Trans Biomed Eng"},{"key":"1942_CR25","doi-asserted-by":"publisher","first-page":"327","DOI":"10.1016\/B978-0-323-39276-1.00039-1","volume-title":"Evidence-based physical diagnosis","author":"S McGee","year":"2018","unstructured":"McGee S. Chapter 39 - auscultation of the heart: general principles. In: McGee S, editor. Evidence-based physical diagnosis. 4th ed. Philadelphia: Elsevier; 2018. p. 327\u201332.","edition":"4"},{"issue":"3","key":"1942_CR26","doi-asserted-by":"publisher","first-page":"872","DOI":"10.1109\/TIP.2012.2219544","volume":"22","author":"W Liu","year":"2013","unstructured":"Liu W, Lin W. Additive white Gaussian noise level estimation in SVD domain for images. IEEE Trans Image Process. 2013;22(3):872\u201383.","journal-title":"IEEE Trans Image Process"},{"key":"1942_CR27","doi-asserted-by":"publisher","unstructured":"Saravanan G, Yamuna G, Nandhini S. Real time implementation of RGB to HSV\/HSI\/HSL and its reverse color space models. In: 2016 International conference on communication and signal processing (ICCSP), Melmaruvathur; 2016. pp. 0462\u20130466. https:\/\/doi.org\/10.1109\/ICCSP.2016.7754179","DOI":"10.1109\/ICCSP.2016.7754179"},{"issue":"6","key":"1942_CR28","doi-asserted-by":"publisher","first-page":"84","DOI":"10.1145\/3065386","volume":"60","author":"A Krizhevsky","year":"2017","unstructured":"Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. Commun ACM. 2017;60(6):84\u201390. https:\/\/doi.org\/10.1145\/3065386.","journal-title":"Commun ACM"},{"key":"1942_CR29","doi-asserted-by":"publisher","unstructured":"Zhang W, Han J. Towards heart sound classification without segmentation using convolutional neural network. In: 2017 Computing in Cardiology (CinC). 2017. p. 1\u20134. https:\/\/doi.org\/10.22489\/CinC.2017.254-164.","DOI":"10.22489\/CinC.2017.254-164"},{"key":"1942_CR30","doi-asserted-by":"crossref","unstructured":"Tschannen M, Kramer T, Marti G, Heinzmann M, Wiatowski T, Heart sound classification using deep structured features. In: 2016 Computing in cardiology conference (CinC), Vancouver, BC; 2016. p. 565\u20138.","DOI":"10.22489\/CinC.2016.162-186"},{"key":"1942_CR31","doi-asserted-by":"publisher","first-page":"1701","DOI":"10.1088\/1361-6579\/aa7623","volume":"38","author":"BM Whitaker","year":"2017","unstructured":"Whitaker BM, et al. Physiol Meas. 2017;38:1701.","journal-title":"Physiol Meas"},{"key":"1942_CR32","doi-asserted-by":"publisher","unstructured":"Shijie J, Ping W, Peiyi J, Siping H, Research on data augmentation for image classification based on convolution neural networks. In: 2017 Chinese automation congress (CAC), Jinan; 2017. p. 4165\u201370. https:\/\/doi.org\/10.1109\/CAC.2017.8243510.","DOI":"10.1109\/CAC.2017.8243510"},{"key":"1942_CR33","doi-asserted-by":"publisher","first-page":"2613","DOI":"10.21437\/Interspeech.2019-2680","volume":"2019","author":"DS Park","year":"2019","unstructured":"Park DS, Chan W, Zhang Y, Chiu C, Zoph B, Cubuk ED, Le QV. SpecAugment: a simple data augmentation method for automatic speech recognition. Proc Interspeech. 2019;2019:2613\u20137. https:\/\/doi.org\/10.21437\/Interspeech.2019-2680.","journal-title":"Proc Interspeech"}],"container-title":["BMC Medical Informatics and Decision Making"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12911-022-01942-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s12911-022-01942-2\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s12911-022-01942-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,8,29]],"date-time":"2022-08-29T08:07:27Z","timestamp":1661760447000},"score":1,"resource":{"primary":{"URL":"https:\/\/bmcmedinformdecismak.biomedcentral.com\/articles\/10.1186\/s12911-022-01942-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,8,29]]},"references-count":33,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2022,12]]}},"alternative-id":["1942"],"URL":"https:\/\/doi.org\/10.1186\/s12911-022-01942-2","relation":{},"ISSN":["1472-6947"],"issn-type":[{"value":"1472-6947","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,8,29]]},"assertion":[{"value":"9 September 2021","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"10 May 2022","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"29 August 2022","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical approval and consent to participate"}},{"value":"Not applicable.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The authors declare that they have no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"226"}}