Memory-efficient low-compute segmentation algorithms for edge ultrasound devices: continuous bladder monitoring

Memory-efficient low-compute segmentation algorithms for edge ultrasound devices: continuous bladder monitoring | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Memory-efficient low-compute segmentation algorithms for edge ultrasound devices: continuous bladder monitoring Zhiye Song, Mercy Asiedu, Shuhang Wang, Qian Li, Arinc Ozturk, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2783298/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 30 Sep, 2023 Read the published version in Scientific Reports → Version 1 posted 9 You are reading this latest preprint version Abstract Post-operative urinary retention is a medical condition where patients cannot urinate despite having a full bladder. Ultrasound imaging of the bladder is intermittently used to estimate urine volume for early diagnosis and management of urine retention. Moreover, the use of bladder ultrasound can reduce the need for an indwelling urinary catheter and the risk of catheter-associated urinary tract infection. Wearable ultrasound devices combined with machine-learning based bladder volume estimation algorithms reduce the burdens of nurses in hospital settings and improve outpatient care. However, existing algorithms are memory and computation intensive, thereby demanding the use of expensive GPUs. In this paper, we develop and validate a low-compute memory-efficient deep learning model for accurate bladder region segmentation and urine volume calculation. B-mode ultrasound bladder images of 360 patients were divided into training and validation sets; another 74 patients were used as the test dataset. Our 1-bit quantized models with 4-bits and 6-bits skip connections achieved an accuracy within 3.8% and 2.6%, respectively, of a full precision state-of-the-art neural network (NN) without any floating-point operations and with an 11.5× and 9.0× reduction in memory requirements to fit under 150 kB. The means and standard deviations of the volume estimation errors, relative to estimates from ground-truth clinician annotations, were 5.0±33 ml and 6.8±29 ml, respectively. This lightweight NN can be easily integrated on the wearable ultrasound device for automated and continuous monitoring of urine volume. Our approach can potentially be extended to other clinical applications such as monitoring blood pressure and fetal heart rate. Health sciences/Health care/Medical imaging/Ultrasonography Physical sciences/Engineering/Biomedical engineering Health sciences/Urology/Bladder Physical sciences/Mathematics and computing/Computer science Physical sciences/Mathematics and computing/Computational science Full Text Additional Declarations Competing interest reported. T.T.P. has no related competing interests. Unrelated disclosures include salary support from the American Roentgen Ray Society, General Electric, the US Department of Defense, and the National Institutes of Health through his institution to support ongoing research projects. He receives royalties from Elsevier Inc. He holds equity and is a consultant for AutonomUS Medical Technologies Inc. He has previously received research support from the Society of Abdominal Radiology and honoraria from the Massachusetts Society of Radiologic Technologists and Zhejiang Medical Association. A.E.S. has served as a compensated consultant for Astra Zeneca, Bracco Diagnostics, Bristol Myers Squibb, General Electric, Gerson Lehman Group, Guidepoint Global Advisors, Supersonic Imagine, Novartis, Pfizer, Philips, Parexel Informatics, and WorldCare Clinical. He holds stock options in Rhino Healthtech Inc. and Ochre Bio Inc. He holds advisory board or committee memberships for General Electric, the Foundation for the National Institutes of Health, and the Sano Center for Computational Personalized Medicine, and is a member of the Board of Governors of the American Institute for Ultrasound in Medicine. His institution has received research grant support and/or equipment for projects that he has led from the Analogic Corporation, Canon, Echosens, General Electric, Hitachi, Philips, Siemens, Supersonic Imagine/Hologic, Toshiba Medical Systems, the US Department of Defense, Fujifilm Healthcare, the Foundation for the National Institutes of Health, Partners Healthcare, Toshiba Medical Systems, and Siemens Medical Systems. He holds equity in Avira Inc., Autonomus Medical Technologies, Inc., Evidence Based Psychology LLC, Klea LLC, Katharos Laboratories LLC, Quantix Bio LLC, and Sonoluminous LLC. He receives royalties from Elsevier Inc. and Katharos Laboratories LLC. A. C. is on the Board of Analog Devices. Other authors declare no competing interests. Supplementary Files SI.pdf Cite Share Download PDF Status: Published Journal Publication published 30 Sep, 2023 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Major revision 30 Jul, 2023 Reviews received at journal 23 Jul, 2023 Reviewers agreed at journal 12 Jul, 2023 Reviewers agreed at journal 10 Jul, 2023 Reviewers invited by journal 07 Jul, 2023 Editor assigned by journal 07 Jul, 2023 Editor invited by journal 14 Apr, 2023 Submission checks completed at journal 14 Apr, 2023 First submitted to journal 05 Apr, 2023 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2783298","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":191833697,"identity":"7f0f0503-f84c-4d85-8b97-8dc10232bc4b","order_by":0,"name":"Zhiye Song","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAqUlEQVRIiWNgGAWjYDACCcYGBoYKBh4wmwQtZxh4eEjQAsSMbQwMxGuRn93cJvFx3jYZewbmg7d5iNHCOOdgm+TMbbeBDmNLtiZKC7NEYps0L1gLj5k0UVrYQFr+zgFp4f9GnBYekBbGBrAtbMRpkZBIbLbsOQbUcpjN2HIOMVrkZ6Q/vPGj5rY9e3vzwxtviNECBCyQ6GAmUjlY7QcSFI+CUTAKRsFIBABDbSka5aex7wAAAABJRU5ErkJggg==","orcid":"","institution":"Massachusetts Institute of Technology","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Zhiye","middleName":"","lastName":"Song","suffix":""},{"id":191833698,"identity":"162781f1-54a1-436c-9551-5f954d5d0f07","order_by":1,"name":"Mercy Asiedu","email":"","orcid":"","institution":"Massachusetts Institute of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Mercy","middleName":"","lastName":"Asiedu","suffix":""},{"id":191833699,"identity":"45e92b35-fe5e-4f44-bc4e-8f31a8251a84","order_by":2,"name":"Shuhang Wang","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Shuhang","middleName":"","lastName":"Wang","suffix":""},{"id":191833702,"identity":"4351a7f6-8c46-478e-9731-d267f81de61d","order_by":3,"name":"Qian Li","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Li","suffix":""},{"id":191833706,"identity":"9c386461-a85e-4cad-96fd-e0d01fe9735e","order_by":4,"name":"Arinc Ozturk","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Arinc","middleName":"","lastName":"Ozturk","suffix":""},{"id":191833707,"identity":"73aad40b-50d1-4801-8002-fb946ba43816","order_by":5,"name":"Vipasha Mittal","email":"","orcid":"","institution":"Massachusetts Institute of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Vipasha","middleName":"","lastName":"Mittal","suffix":""},{"id":191833708,"identity":"0084c8cf-033d-487f-99c1-b9fdc1253803","order_by":6,"name":"Scott Schoen Jr.","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Scott","middleName":"","lastName":"Schoen","suffix":"Jr."},{"id":191833709,"identity":"01681ff6-a83d-493f-b89f-da7da54d276d","order_by":7,"name":"Srinath Ramaswamy","email":"","orcid":"","institution":"Texas Instruments (United States)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Srinath","middleName":"","lastName":"Ramaswamy","suffix":""},{"id":191833710,"identity":"5b77153a-c2bf-48ff-a92f-e8b6dee077b6","order_by":8,"name":"Theodore T. Pierce","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Theodore","middleName":"T.","lastName":"Pierce","suffix":""},{"id":191833711,"identity":"f9df15a8-5c39-4bfe-9bb7-8f6494529ee7","order_by":9,"name":"Anthony E. Samir","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Anthony","middleName":"E.","lastName":"Samir","suffix":""},{"id":191833712,"identity":"a214adad-1202-4d2d-9bc9-be4c1fcb6e40","order_by":10,"name":"Yonina C. Eldar","email":"","orcid":"","institution":"Weizmann Institute of Science","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Yonina","middleName":"C.","lastName":"Eldar","suffix":""},{"id":191833713,"identity":"817d187e-82e2-4776-8f4b-d9e5d8d03aa9","order_by":11,"name":"Anantha Chandrakasan","email":"","orcid":"","institution":"Massachusetts Institute of Technology","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Anantha","middleName":"","lastName":"Chandrakasan","suffix":""},{"id":191833714,"identity":"21919a30-e782-4833-842b-3eb2822cce1e","order_by":12,"name":"Viksit Kumar","email":"","orcid":"","institution":"Massachusetts General Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Viksit","middleName":"","lastName":"Kumar","suffix":""}],"badges":[],"createdAt":"2023-04-06 02:44:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2783298/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2783298/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-023-42000-9","type":"published","date":"2023-09-30T15:01:32+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":43974752,"identity":"5d9f950f-3fbe-4493-a569-82f24c6110ce","added_by":"auto","created_at":"2023-10-02 15:09:56","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1119756,"visible":true,"origin":"","legend":"","description":"","filename":"Memoryefficientlowcomputesegmentationalgorithmsforedgeultrasounddevicescontinuousbladdermonitoring.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2783298/v1_covered_b306a092-ead9-4c0b-9b75-bc0a0eae03ee.pdf"},{"id":35920951,"identity":"bf294243-6ed5-4627-b15a-31754c261062","added_by":"auto","created_at":"2023-04-18 08:25:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":262660,"visible":true,"origin":"","legend":"","description":"","filename":"SI.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2783298/v1/94770a1594404d303da9be5d.pdf"}],"financialInterests":"Competing interest reported. T.T.P. has no related competing interests. Unrelated disclosures include salary support from the American Roentgen Ray Society, General Electric, the US Department of Defense, and the National Institutes of Health through his institution to support ongoing research projects. He receives royalties from Elsevier Inc. He holds equity and is a consultant for AutonomUS Medical Technologies Inc. He has previously received research support from the Society of Abdominal Radiology and honoraria from the Massachusetts Society of Radiologic Technologists and Zhejiang Medical Association.\n\nA.E.S. has served as a compensated consultant for Astra Zeneca, Bracco Diagnostics, Bristol Myers Squibb, General Electric, Gerson Lehman Group, Guidepoint Global Advisors, Supersonic Imagine, Novartis, Pfizer, Philips, Parexel Informatics, and WorldCare Clinical. He holds stock options in Rhino Healthtech Inc. and Ochre Bio Inc. He holds advisory board or committee memberships for General Electric, the Foundation for the National Institutes of Health, and the Sano Center for Computational Personalized Medicine, and is a member of the Board of Governors of the American Institute for Ultrasound in Medicine. His institution has received research grant support and/or equipment for projects that he has led from the Analogic Corporation, Canon, Echosens, General Electric, Hitachi, Philips, Siemens, Supersonic Imagine/Hologic, Toshiba Medical Systems, the US Department of Defense, Fujifilm Healthcare, the Foundation for the National Institutes of Health, Partners Healthcare, Toshiba Medical Systems, and Siemens Medical Systems. He holds equity in Avira Inc., Autonomus Medical Technologies, Inc., Evidence Based Psychology LLC, Klea LLC, Katharos Laboratories LLC, Quantix Bio LLC, and Sonoluminous LLC. He receives royalties from Elsevier Inc. and Katharos Laboratories LLC.\n\nA. C. is on the Board of Analog Devices.\n\nOther authors declare no competing interests.","formattedTitle":"Memory-efficient low-compute segmentation algorithms for edge ultrasound devices: continuous bladder monitoring","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-2783298/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2783298/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Post-operative urinary retention is a medical condition where patients cannot urinate despite having a full bladder. Ultrasound imaging of the bladder is intermittently used to estimate urine volume for early diagnosis and management of urine retention. Moreover, the use of bladder ultrasound can reduce the need for an indwelling urinary catheter and the risk of catheter-associated urinary tract infection. Wearable ultrasound devices combined with machine-learning based bladder volume estimation algorithms reduce the burdens of nurses in hospital settings and improve outpatient care. However, existing algorithms are memory and computation intensive, thereby demanding the use of expensive GPUs. In this paper, we develop and validate a low-compute memory-efficient deep learning model for accurate bladder region segmentation and urine volume calculation. B-mode ultrasound bladder images of 360 patients were divided into training and validation sets; another 74 patients were used as the test dataset. Our 1-bit quantized models with 4-bits and 6-bits skip connections achieved an accuracy within 3.8% and 2.6%, respectively, of a full precision state-of-the-art neural network (NN) without any floating-point operations and with an 11.5× and 9.0× reduction in memory requirements to fit under 150 kB. The means and standard deviations of the volume estimation errors, relative to estimates from ground-truth clinician annotations, were 5.0±33 ml and 6.8±29 ml, respectively. This lightweight NN can be easily integrated on the wearable ultrasound device for automated and continuous monitoring of urine volume. Our approach can potentially be extended to other clinical applications such as monitoring blood pressure and fetal heart rate.","manuscriptTitle":"Memory-efficient low-compute segmentation algorithms for edge ultrasound devices: continuous bladder monitoring","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-04-18 08:25:42","doi":"10.21203/rs.3.rs-2783298/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revision","date":"2023-07-30T10:23:34+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2023-07-23T07:51:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"53094d0f-33af-4f82-af1c-0fad00637ed9","date":"2023-07-12T21:01:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"f6e00652-96f5-4d25-b6c3-21a36e6add30","date":"2023-07-11T01:52:32+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2023-07-07T15:33:32+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2023-07-07T05:59:21+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2023-04-14T09:05:26+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2023-04-14T08:57:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2023-04-06T02:43:12+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c0e470b4-3c6e-4a2f-9247-e759656f237c","owner":[],"postedDate":"April 18th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":20705815,"name":"Health sciences/Health care/Medical imaging/Ultrasonography"},{"id":20705816,"name":"Physical sciences/Engineering/Biomedical engineering"},{"id":20705817,"name":"Health sciences/Urology/Bladder"},{"id":20705818,"name":"Physical sciences/Mathematics and computing/Computer science"},{"id":20705819,"name":"Physical sciences/Mathematics and computing/Computational science"}],"tags":[],"updatedAt":"2023-10-02T15:07:55+00:00","versionOfRecord":{"articleIdentity":"rs-2783298","link":"https://doi.org/10.1038/s41598-023-42000-9","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2023-09-30 15:01:32","publishedOnDateReadable":"September 30th, 2023"},"versionCreatedAt":"2023-04-18 08:25:42","video":"","vorDoi":"10.1038/s41598-023-42000-9","vorDoiUrl":"https://doi.org/10.1038/s41598-023-42000-9","workflowStages":[]},"version":"v1","identity":"rs-2783298","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2783298","identity":"rs-2783298","version":["v1"]},"buildId":"7rjqhiLT3MXkJMwkYKINL","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}