{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,21]],"date-time":"2026-09-21T16:06:51Z","timestamp":1790006811726,"version":"4.0.1"},"reference-count":35,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,2,1]],"date-time":"2025-02-01T00:00:00Z","timestamp":1738368000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,2,1]],"date-time":"2025-02-01T00:00:00Z","timestamp":1738368000000},"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 biomed eng"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Aim<\/jats:title>\n                    <jats:p>The aim of this study is to apply a novel hybrid framework incorporating a Vision Transformer (ViT) and bidirectional long short-term memory (Bi-LSTM) model for classifying physical activity intensity (PAI) in adults using gravity-based acceleration. Additionally, it further investigates how PAI and temporal window (TW) impacts the model\u2019 s accuracy.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Method<\/jats:title>\n                    <jats:p>This research used the Capture-24 dataset, consisting of raw accelerometer data from 151 participants aged 18 to 91. Gravity-based acceleration was utilised to generate images encoding various PAIs. These images were subsequently analysed using the ViT-BiLSTM model, with results presented in confusion matrices and compared with baseline models. The model\u2019s robustness was evaluated through temporal stability testing and examination of accuracy and loss curves.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Result<\/jats:title>\n                    <jats:p>\n                      The ViT-BiLSTM model excelled in PAI classification task, achieving an overall accuracy of 98.5%\u2009\u00b1\u20091.48% across five TWs-98.7% for 1s, 98.1% for 5s, 98.2% for 10s, 99% for 15s, and 98.65% for 30s of TW. The model consistently exhibited superior accuracy in predicting sedentary (98.9%\u2009\u00b1\u20091%) compared to light physical activity (98.2%\u2009\u00b1\u20092%) and moderate-to-vigorous physical activity (98.2%\u2009\u00b1\u20093%). ANOVA showed no significant accuracy variation across PAIs (F\u2009=\u20092.18,\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.13) and TW (F\u2009=\u20090.52,\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.72). Accuracy and loss curves show the model consistently improves its performance across epochs, demonstrating its excellent robustness.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This study demonstrates the ViT-BiLSTM model\u2019s efficacy in classifying PAI using gravity-based acceleration, with performance remaining consistent across diverse TWs and intensities. However, PAI and TW could result in slight variations in the model\u2019s performance. Future research should concern and investigate the impact of gravity-based acceleration on PAI thresholds, which may influence model\u2019s robustness and reliability.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s42490-025-00088-2","type":"journal-article","created":{"date-parts":[[2025,1,31]],"date-time":"2025-01-31T18:02:48Z","timestamp":1738346568000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":9,"title":["A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration"],"prefix":"10.1186","volume":"7","author":[{"given":"Lin","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zizhang","family":"Luo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tianle","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,2,1]]},"reference":[{"issue":"4-5","key":"88_CR1","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1080\/10408399309527627","volume":"33","author":"RR Pate","year":"1993","unstructured":"Pate RR. Physical activity assessment in children and adolescents. Crit Rev Food Sci Nutr. 1993;33(4-5):321\u201326.","journal-title":"Crit Rev Food Sci Nutr"},{"issue":"3","key":"88_CR2","doi-asserted-by":"publisher","first-page":"252","DOI":"10.1123\/pes.19.3.252","volume":"19","author":"AV Rowlands","year":"2007","unstructured":"Rowlands AV. Accelerometer assessment of physical activity in children: an update. Pediatr Exerc Sci. 2007;19(3):252\u201366.","journal-title":"Pediatr Exerc Sci"},{"issue":"4","key":"88_CR3","doi-asserted-by":"publisher","first-page":"450","DOI":"10.1123\/pes.2018-0201","volume":"30","author":"AV Rowlands","year":"2018","unstructured":"Rowlands AV. Moving forward with accelerometer-assessed physical activity: two strategies to ensure meaningful, interpretable, and comparable measures. Pediatr Exerc Sci. 2018;30(4):450\u201356.","journal-title":"Pediatr Exerc Sci"},{"key":"88_CR4","doi-asserted-by":"publisher","first-page":"1323","DOI":"10.1249\/MSS.0000000000001561","volume":"50","author":"AV Rowlands","year":"2018","unstructured":"Rowlands AV, et al. Beyond cut points: accelerometer metrics that capture the physical activity profile. Med Sci Sports Exerc 2018;50:1323\u201332.","journal-title":"Med Sci Sports Exerc"},{"key":"88_CR5","doi-asserted-by":"publisher","first-page":"2159-2167","DOI":"10.1080\/02640414.2019.1624313","volume":"37","author":"SJ Fairclough","year":"2019","unstructured":"Fairclough SJ, et al. Average acceleration and intensity gradient of primary school children and associations with indicators of health and well-being. J Sports Sci 2019;37:2159-2167.","journal-title":"J Sports Sci"},{"key":"88_CR6","doi-asserted-by":"publisher","first-page":"e800","DOI":"10.1016\/S2468-2667(23)00183-4","volume":"8","author":"MN Ahmadi","year":"2023","unstructured":"Ahmadi MN, et al. Brief bouts of device-measured intermittent lifestyle physical activity and its association with major adverse cardiovascular events and mortality in people who do not exercise: a prospective cohort study. Lancet Public Health 2023;8:e800\u2013e810.","journal-title":"Lancet Public Health"},{"issue":"1","key":"88_CR7","doi-asserted-by":"publisher","first-page":"115","DOI":"10.1111\/pedi.13288","volume":"23","author":"Z Marshall","year":"2022","unstructured":"Marshall Z, Mackintosh K, Gregory J, McNarry M. Using compositional analysis to explore the relationship between physical activity and cardiovascular health in children and adolescents with and without type 1 diabetes. Pediatr Diabetes. 2022;23(1):115\u201325.","journal-title":"Pediatr Diabetes"},{"key":"88_CR8","doi-asserted-by":"publisher","first-page":"376","DOI":"10.1136\/bjsports-2020-103604","volume":"56","author":"JH Migueles","year":"2022","unstructured":"Migueles JH, et al. GRANADA consensus on analytical approaches to assess associations with accelerometer-determined physical behaviours (physical activity, sedentary behaviour and sleep) in epidemiological studies. Br J Sports Med 2022;56:376\u201384.","journal-title":"Br J Sports Med"},{"key":"88_CR9","doi-asserted-by":"publisher","first-page":"e0299295","DOI":"10.1371\/journal.pone.0299295","volume":"19","author":"JYY Ng","year":"2024","unstructured":"Ng JYY, et al. Development of a multi-wear-site, deep learning-based physical activity intensity classification algorithm using raw acceleration data. PLoS One 2024;19:e0299295.","journal-title":"PLoS One"},{"key":"88_CR10","doi-asserted-by":"publisher","first-page":"86","DOI":"10.1038\/s41746-024-01065-0","volume":"7","author":"H Yuan","year":"2024","unstructured":"Yuan H, et al. Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality. NPJ Digit Med 2024;7:86.","journal-title":"NPJ Digit Med"},{"issue":"4","key":"88_CR11","doi-asserted-by":"publisher","first-page":"1324","DOI":"10.1152\/japplphysiol.00818.2005","volume":"100","author":"SE Crouter","year":"2006","unstructured":"Crouter SE, Clowers KG, Bassett Jr DR. A novel method for using accelerometer data to predict energy expenditure. J Appl Physiol (1985). 2006;100(4):1324\u201331.","journal-title":"J Appl Physiol (1985)"},{"issue":"5","key":"88_CR12","doi-asserted-by":"publisher","first-page":"777","DOI":"10.1097\/00005768-199805000-00021","volume":"30","author":"PS Freedson","year":"1998","unstructured":"Freedson PS, Melanson E, Sirard J. Calibration of the computer science and applications, inc. accelerometer. Med Sci Sports Exerc. 1998;30(5):777\u201381.","journal-title":"Med Sci Sports Exerc"},{"key":"88_CR13","doi-asserted-by":"publisher","first-page":"E146","DOI":"10.5888\/pcd13.160159","volume":"13","author":"WR Boyer","year":"2016","unstructured":"Boyer WR. Accelerometer-derived total activity counts, bouted minutes of moderate to vigorous activity, and insulin resistance: NHANES 2003\u20132006. Prev Chronic Dis. 2016;13:E146.","journal-title":"Prev Chronic Dis"},{"key":"88_CR14","doi-asserted-by":"publisher","first-page":"1132","DOI":"10.1016\/j.jsams.2019.06.016","volume":"22","author":"AV Rowlands","year":"2019","unstructured":"Rowlands AV, et al. A data-driven, meaningful, easy to interpret, standardised accelerometer outcome variable for global surveillance. J Sci Med Sport 2019;22:1132\u201338.","journal-title":"J Sci Med Sport"},{"key":"88_CR15","doi-asserted-by":"publisher","first-page":"105004","DOI":"10.1016\/j.ijmedinf.2023.105004","volume":"172","author":"V Farrahi","year":"2023","unstructured":"Farrahi V, Muhammad U, Rostami M, Oussalah M. AccNet24: a deep learning framework for classifying 24-hour activity behaviours from wrist-worn accelerometer data under free-living environments. Int J Med Inform. 2023;172:105004.","journal-title":"Int J Med Inform"},{"key":"88_CR16","doi-asserted-by":"publisher","first-page":"683","DOI":"10.1080\/02640414.2020.1841394","volume":"39","author":"R Nawaratne","year":"2021","unstructured":"Nawaratne R, et al. Deep learning to predict energy expenditure and activity intensity in free living conditions using wrist-specific accelerometry. J Sports Sci 2021;39:683\u201390.","journal-title":"J Sports Sci"},{"issue":"4","key":"88_CR17","doi-asserted-by":"publisher","first-page":"e0000220","DOI":"10.1371\/journal.pdig.0000220","volume":"2","author":"CB Thornton","year":"2023","unstructured":"Thornton CB, Kolehmainen N, Nazarpour K. Using unsupervised machine learning to quantify physical activity from accelerometry in a diverse and rapidly changing population. PLOS Digit Health. 2023;2(4):e0000220.","journal-title":"PLOS Digit Health"},{"key":"88_CR18","doi-asserted-by":"publisher","first-page":"e0208692","DOI":"10.1371\/journal.pone.0208692","volume":"14","author":"D van Kuppevelt","year":"2019","unstructured":"van Kuppevelt D, et al. Segmenting accelerometer data from daily life with unsupervised machine learning. PLoS One 2019;14:e0208692.","journal-title":"PLoS One"},{"key":"88_CR19","doi-asserted-by":"publisher","first-page":"e241","DOI":"10.1016\/j.scispo.2019.02.006","volume":"34","author":"R Muazu Musa","year":"2019","unstructured":"Muazu Musa R, et al. The application of Artificial Neural Network and k-Nearest Neighbour classification models in the scouting of high-performance archers from a selected fitness and motor skill performance parameters. Sci Sports 2019;34:e241\u2013e249.","journal-title":"Sci Sports"},{"key":"88_CR20","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1038\/s41746-024-01062-3","volume":"7","author":"H Yuan","year":"2024","unstructured":"Yuan H, et al. Self-supervised learning for human activity recognition using 700,000 person-days of wearable data. NPJ Digit Med 2024;7:91.","journal-title":"NPJ Digit Med"},{"key":"88_CR21","volume-title":"Physical Activity Intensity Classification Using a Convolutional Neural Network and Wearable Accelerometer","author":"A Widianto","year":"2019","unstructured":"Widianto A, et al. Physical Activity Intensity Classification Using a Convolutional Neural Network and Wearable Accelerometer. 2019."},{"key":"88_CR22","doi-asserted-by":"publisher","first-page":"45","DOI":"10.1016\/j.gaitpost.2021.06.017","volume":"89","author":"J Chong","year":"2021","unstructured":"Chong J, et al. Machine-learning models for activity class prediction: a comparative study of feature selection and classification algorithms. Gait Posture. 2021;89:45\u201353.","journal-title":"Gait Posture"},{"key":"88_CR23","doi-asserted-by":"publisher","first-page":"118484","DOI":"10.1016\/j.eswa.2022.118484","volume":"210","author":"KS Tan","year":"2022","unstructured":"Tan KS, Lim KM, Lee CP, Kwek LC. Bidirectional long short-term memory with temporal dense sampling for human action recognition. Expert Syst Appl. 2022;210:118484.","journal-title":"Expert Syst Appl"},{"key":"88_CR24","doi-asserted-by":"publisher","first-page":"121978","DOI":"10.1016\/j.eswa.2023.121978","volume":"238","author":"Y Zhao","year":"2024","unstructured":"Zhao Y, et al. Image expression of time series data of wearable IMU sensor and fusion classification of gymnastics action. Expert Syst Appl. 2024;238:121978.","journal-title":"Expert Syst Appl"},{"key":"88_CR25","doi-asserted-by":"publisher","first-page":"11654","DOI":"10.1038\/s41598-022-15245-z","volume":"12","author":"D Vela","year":"2022","unstructured":"Vela D, et al. Temporal quality degradation in AI models. Sci Rep 2022;12:11654.","journal-title":"Sci Rep"},{"key":"88_CR26","doi-asserted-by":"publisher","first-page":"127","DOI":"10.1186\/s12966-023-01526-x","volume":"20","author":"A Miatke","year":"2023","unstructured":"Miatke A, et al. The association between reallocations of time and health using compositional data analysis: a systematic scoping review with an interactive data exploration interface. Int J Behav Nutr Phys Act 2023;20:127.","journal-title":"Int J Behav Nutr Phys Act"},{"key":"88_CR27","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1186\/s40798-019-0225-9","volume":"5","author":"AV Rowlands","year":"2019","unstructured":"Rowlands AV, et al. Enhancing the value of accelerometer-assessed physical activity: meaningful visual comparisons of data-driven translational accelerometer metrics. Sports Med Open 2019;5:47.","journal-title":"Sports Med Open"},{"key":"88_CR28","doi-asserted-by":"publisher","first-page":"318","DOI":"10.1177\/0081175019884591","volume":"50","author":"J Gershuny","year":"2020","unstructured":"Gershuny J, et al. Testing self-report time-use diaries against objective instruments in real time. Socioll Methodol 2020;50:318\u201349.","journal-title":"Socioll Methodol"},{"key":"88_CR29","doi-asserted-by":"publisher","first-page":"318","DOI":"10.1177\/0081175019884591","volume":"50","author":"J Gershuny","year":"2019","unstructured":"Gershuny J, et al. Testing self-report time-use diaries against objective instruments in real time. Socioll Methodol 2019;50:318\u201349.","journal-title":"Socioll Methodol"},{"key":"88_CR30","doi-asserted-by":"publisher","first-page":"e61691","DOI":"10.1371\/journal.pone.0061691","volume":"8","author":"VT van Hees","year":"2013","unstructured":"van Hees VT, et al. Separating movement and gravity components in an acceleration signal and implications for the assessment of human daily physical activity. PLoS One 2013;8:e61691.","journal-title":"PLoS One"},{"issue":"9","key":"88_CR31","doi-asserted-by":"publisher","first-page":"1816","DOI":"10.1249\/MSS.0000000000000289","volume":"46","author":"M Hildebrand","year":"2014","unstructured":"Hildebrand M, van Hees VT, Hansen BH, Ekelund U. Age group comparability of raw accelerometer output from wrist- and hip-worn monitors. Med Sci Sports Exerc. 2014;46(9):1816\u201324.","journal-title":"Med Sci Sports Exerc"},{"key":"88_CR32","doi-asserted-by":"publisher","first-page":"7961","DOI":"10.1038\/s41598-018-26174-1","volume":"8","author":"M Willetts","year":"2018","unstructured":"Willetts M, et al. Statistical machine learning of sleep and physical activity phenotypes from sensor data in 96,220 UK Biobank participants. Sci Rep 2018;8:7961.","journal-title":"Sci Rep"},{"key":"88_CR33","doi-asserted-by":"publisher","first-page":"1008","DOI":"10.1136\/bjsports-2021-104050","volume":"56","author":"R Walmsley","year":"2022","unstructured":"Walmsley R, et al. Reallocation of time between device-measured movement behaviours and risk of incident cardiovascular disease. Br J Sports Med 2022;56:1008\u201317.","journal-title":"Br J Sports Med"},{"issue":"2","key":"88_CR34","first-page":"126","volume":"100","author":"CJ Caspersen","year":"1985","unstructured":"Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep. 1985;100(2):126.","journal-title":"Public Health Rep"},{"key":"88_CR35","doi-asserted-by":"publisher","first-page":"234","DOI":"10.1016\/j.jshs.2021.03.004","volume":"11","author":"S Verswijveren","year":"2022","unstructured":"Verswijveren S, et al. Using compositional data analysis to explore accumulation of sedentary behavior, physical activity and youth health. J Sport Health Sci 2022;11:234\u201343.","journal-title":"J Sport Health Sci"}],"container-title":["BMC Biomedical Engineering"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s42490-025-00088-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s42490-025-00088-2\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s42490-025-00088-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,31]],"date-time":"2025-01-31T18:02:51Z","timestamp":1738346571000},"score":1,"resource":{"primary":{"URL":"https:\/\/bmcbiomedeng.biomedcentral.com\/articles\/10.1186\/s42490-025-00088-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,1]]},"references-count":35,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["88"],"URL":"https:\/\/doi.org\/10.1186\/s42490-025-00088-2","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-4696057\/v1","asserted-by":"object"}]},"ISSN":["2524-4426"],"issn-type":[{"value":"2524-4426","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,1]]},"assertion":[{"value":"6 July 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 January 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"1 February 2025","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, this study used the open dataset Capture-24, which contains Axivity AX3 wrist-worn activity tracker data collected from 151 participants in the Oxfordshire area during 2014-2016. The the link of the dataset in the data availability:\n                      \n                      . For detailed data collection protocols, see the following references: Gershuny et al. (2020), Sociological Methodology, doi:\n                      \n                      ; Willetts et al. (2018), Scientific Reports, 8(1):7961.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"Not applicable.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"2"}}