A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration

Wang, Lin; Luo, Zizhang; Zhang, Tianle

doi:10.1186/s42490-025-00088-2

Cited by 8Context-Aware Activity Recognition SystemsPhysical Activity and HealthTime Series Analysis and Forecasting

A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration

Open access

Wang, Lin, Luo, Zizhang, Zhang, Tianle·2025·BMC Biomedical Engineering

DOI 10.1186/s42490-025-00088-2PMC PMC11786420

Published version of a preprint

Open this paper in CoChat — free account Open PDF

Abstract

Aim: 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’ s accuracy. Method: 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’s robustness was evaluated through temporal stability testing and examination of accuracy and loss curves. Result: The ViT-BiLSTM model excelled in PAI classification task, achieving an overall accuracy of 98.5% ± 1.48% 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% ± 1%) compared to light physical activity (98.2% ± 2%) and moderate-to-vigorous physical activity (98.2% ± 3%). ANOVA showed no significant accuracy variation across PAIs (F = 2.18, p = 0.13) and TW (F = 0.52, p = 0.72). Accuracy and loss curves show the model consistently improves its performance across epochs, demonstrating its excellent robustness. Conclusion: This study demonstrates the ViT-BiLSTM model’s 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’s performance. Future research should concern and investigate the impact of gravity-based acceleration on PAI thresholds, which may influence model’s robustness and reliability.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

Integrity signals

Registry

Crossref · 10.1186/s42490-025-00088-2

Cites

36 of 36 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research.
(1985) · Europe PMC (PubMed Central) · Cited by 9,143 · see citing works
Calibration of the Computer Science and Applications, Inc. accelerometer
FREEDSON, P. S., MELANSON, E., & SIRARD, J. (1998) · Medicine & Science in Sports & Exercise · doi:10.1097/00005768-199805000-00021 · Cited by 3,825 · see citing works
Age Group Comparability of Raw Accelerometer Output from Wrist- and Hip-Worn Monitors
HILDEBRAND, M., VAN HEES, V. T., HANSEN, B. H., & EKELUND, U. (2014) · Medicine & Science in Sports & Exercise · doi:10.1249/mss.0000000000000289 · Cited by 1,043 · see citing works
Separating Movement and Gravity Components in an Acceleration Signal and Implications for the Assessment of Human Daily Physical Activity
van Hees, V. T., Gorzelniak, L., Dean‐Leon, E., Eder, M., Pias, M. R., Taherian, S., Ekelund, U., Renström, F., Franks, P. W., Horsch, A., et al. (2013) · PLoS ONE · doi:10.1371/journal.pone.0061691 · Cited by 937 · see citing works
A novel method for using accelerometer data to predict energy expenditure
Crouter, S. E., Clowers, K. G., & Bassett, D. R. (2005) · Journal of Applied Physiology · doi:10.1152/japplphysiol.00818.2005 · Cited by 455 · see citing works
Accelerometer Assessment of Physical Activity in Children: An Update
Rowlands, A. V. (2007) · Pediatric Exercise Science · doi:10.1123/pes.19.3.252 · Cited by 342 · see citing works
Statistical machine learning of sleep and physical activity phenotypes from sensor data in 96,220 UK Biobank participants
Willetts, M., Hollowell, S., Aslett, L., Holmes, C., & Doherty, A. (2018) · Scientific Reports · doi:10.1038/s41598-018-26174-1 · Cited by 250 · see citing works
Beyond Cut Points: Accelerometer Metrics that Capture the Physical Activity Profile
Rowlands, A. V., Edwardson, C. L., Davies, M. J., Khunti, K. K., Harrington, D. M., & Yates, T. E. (2018) · Medicine & Science in Sports & Exercise · doi:10.1249/mss.0000000000001561 · Cited by 244 · see citing works
Reallocation of time between device-measured movement behaviours and risk of incident cardiovascular disease
Walmsley, R., Chan, S., Smith-Byrne, K., Ramakrishnan, R., Woodward, M., Rahimi, K., Dwyer, T., Bennett, D., & Doherty, A. (2022) · British Journal of Sports Medicine · doi:10.1136/bjsports-2021-104050 · Cited by 238 · see citing works
Temporal quality degradation in AI models
Jarquin, D. V., Sharp, A., Zhang, R., Nguyen, T., Hoang, A. T., & Pianykh, O. S. (2022) · Scientific Reports · doi:10.1038/s41598-022-15245-z · Cited by 187 · see citing works
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
Ahmadi, M. N., Hamer, M., Gill, J. M. R., Murphy, M., Sanders, J. P., Doherty, A., & Stamatakis, E. (2023) · The Lancet Public Health · doi:10.1016/s2468-2667(23)00183-4 · Cited by 172 · see citing works
Self-supervised learning for human activity recognition using 700,000 person-days of wearable data
Yuan, H., Chan, S., Creagh, A. P., Tong, C., Acquah, A., Clifton, D. A., & Doherty, A. (2024) · npj Digital Medicine · doi:10.1038/s41746-024-01062-3 · Cited by 166 · see citing works
GRANADA consensus on analytical approaches to assess associations with accelerometer-determined physical behaviours (physical activity, sedentary behaviour and sleep) in epidemiological studies
Migueles, J. H., Aadland, E., Andersen, L. B., Brønd, J. C., CHASTIN, S. F. M., Hansen, B. H., Konstabel, K., Kvalheim, O. M., McGregor, D., Rowlands, A. V., et al. (2021) · British Journal of Sports Medicine · doi:10.1136/bjsports-2020-103604 · Cited by 147 · see citing works
Physical activity assessment in children and adolescents
Pate, R. R. (1993) · Critical Reviews in Food Science and Nutrition · doi:10.1080/10408399309527627 · Cited by 129 · see citing works
Testing Self-Report Time-Use Diaries against Objective Instruments in Real Time
Gershuny, J., Harms, T., Doherty, A., Thomas, E., Milton, K., Kelly, P., & Foster, C. (2020) · Sociological Methodology · doi:10.1177/0081175019884591 · Cited by 124 · see citing works
Enhancing the value of accelerometer-assessed physical activity: meaningful visual comparisons of data-driven translational accelerometer metrics
Rowlands, A. V., Dawkins, N. P., Maylor, B. D., Edwardson, C. L., Fairclough, S. J., Davies, M. J., Harrington, D. M., Khunti, K. K., & Yates, T. (2019) · Sports Medicine - Open · doi:10.1186/s40798-019-0225-9 · Cited by 89 · see citing works
Moving forward with accelerometer-assessed physical activity: two strategies to ensure meaningful, interpretable, and comparable measures
Rowlands, A. V. (2018) · Pediatric Exercise Science · doi:10.1123/pes.2018-0201 · Cited by 87 · see citing works
The association between reallocations of time and health using compositional data analysis: a systematic scoping review with an interactive data exploration interface
Miatke, A., Olds, T., Maher, C. A., Fraysse, F., Mellow, M. L., Smith, A. E., Pedišić, Ž., Grgić, J., & Dumuid, D. (2023) · International Journal of Behavioral Nutrition and Physical Activity · doi:10.1186/s12966-023-01526-x · Cited by 84 · see citing works
Validity of estimating physical activity intensity using a triaxial accelerometer in healthy adults and older adults
Nagayoshi, S., Oshima, Y., Ando, T., Aoyama, T., Nakae, S., Usui, C., Kumagai, S., & Tanaka, S. (2019) · BMJ Open Sport & Exercise Medicine · doi:10.1136/bmjsem-2019-000592 · Cited by 78 · see citing works
A data-driven, meaningful, easy to interpret, standardised accelerometer outcome variable for global surveillance
Rowlands, A. V., Sherar, L. B., Fairclough, S. J., Yates, T. E., Edwardson, C. L., Harrington, D. M., Davies, M. J., Munir, F., Khunti, K. K., & Stiles, V. H. (2019) · Journal of science and medicine in sport · doi:10.1016/j.jsams.2019.06.016 · Cited by 75 · see citing works
Machine-learning models for activity class prediction: A comparative study of feature selection and classification algorithms
Chong, J. L., Tjurin, P., Niemelä, M., Jämsä, T. J., & Farrahi, V. (2021) · Gait & Posture · doi:10.1016/j.gaitpost.2021.06.017 · Cited by 71 · see citing works
Average acceleration and intensity gradient of primary school children and associations with indicators of health and well-being
Fairclough, S. J., Taylor, S. L., Rowlands, A. V., Boddy, L. M., & Noonan, R. J. (2019) · Journal of Sports Sciences · doi:10.1080/02640414.2019.1624313 · Cited by 70 · see citing works
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
Musa, R. M., Majeed, A. P. P. A., Taha, Z., Abdullah, M. R., Maliki, A. B. H. M., & Kosni, N. A. (2019) · Science & Sports · doi:10.1016/j.scispo.2019.02.006 · Cited by 62 · see citing works
Image expression of time series data of wearable IMU sensor and fusion classification of gymnastics action
Zhao, Y., Dong, F., Sun, T., Ju, Z., Yang, L., Shan, P., Li, L., Lv, X., & Lian, C. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121978 · Cited by 49 · see citing works
Bidirectional Long Short-Term Memory with Temporal Dense Sampling for human action recognition
Tan, K. S., Lim, K. M., Lee, C. P., & Kwek, L. C. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.118484 · Cited by 48 · see citing works
Segmenting accelerometer data from daily life with unsupervised machine learning
van Kuppevelt, D., Heywood, J., Hamer, M., Sabia, S., Fitzsimons, E., & van Hees, V. (2019) · PLOS ONE · doi:10.1371/journal.pone.0208692 · Cited by 41 · see citing works
Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality
Yuan, H., Plekhanova, T., Walmsley, R., Reynolds, A. C., Maddison, K. J., Bucan, M., Gehrman, P., Rowlands, A., Ray, D. W., Bennett, D., et al. (2024) · npj Digital Medicine · doi:10.1038/s41746-024-01065-0 · Cited by 40 · see citing works
Using compositional data analysis to explore accumulation of sedentary behavior, physical activity and youth health
Verswijveren, S. J. J. M., Lamb, K. E., Fernández, J. A. M., WINKLER, E. A. H., Leech, R. M., Timperio, A. F., Salmon, J., Daly, R. M., Cerin, E., Dunstan, D. W., et al. (2021) · Journal of sport and health science/Journal of Sport and Health Science · doi:10.1016/j.jshs.2021.03.004 · Cited by 39 · see citing works
Accelerometer-Derived Total Activity Counts, Bouted Minutes of Moderate to Vigorous Activity, and Insulin Resistance: NHANES 2003–2006
Boyer, W. R., Wolff‐Hughes, D. L., Bassett, D. R., Churilla, J. R., & Fitzhugh, E. C. (2016) · Preventing Chronic Disease · doi:10.5888/pcd13.160159 · Cited by 29 · see citing works
Using compositional analysis to explore the relationship between physical activity and cardiovascular health in children and adolescents with and without type 1 diabetes
Marshall, Z. A., Mackintosh, K. A., Gregory, J. W., & McNarry, M. A. (2021) · Pediatric Diabetes · doi:10.1111/pedi.13288 · Cited by 26 · see citing works
AccNet24: A deep learning framework for classifying 24-hour activity behaviours from wrist-worn accelerometer data under free-living environments
Farrahi, V., Muhammad, U., Rostami, M., & Oussalah, M. (2023) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2023.105004 · Cited by 19 · see citing works
Using unsupervised machine learning to quantify physical activity from accelerometry in a diverse and rapidly changing population
Thornton, C., Kolehmainen, N., & Nazarpour, K. (2023) · PLOS Digital Health · doi:10.1371/journal.pdig.0000220 · Cited by 15 · see citing works
Development of a multi-wear-site, deep learning-based physical activity intensity classification algorithm using raw acceleration data
Ng, J. Y. Y., Zhang, J. H., Hui, S. S., Jiang, G., Yau, F., Cheng, J. S.-C., & Ha, A. S. C. (2024) · PLoS ONE · doi:10.1371/journal.pone.0299295 · Cited by 13 · see citing works
Deep Learning to Predict Energy Expenditure and Activity Intensity in Free Living Conditions using Wrist-specific Accelerometry
Nawaratne, R., Alahakoon, D., Silva, D. D., O’Halloran, P., Montoye, A. H. K., Staley, K., Nicholson, M., & Kingsley, M. (2020) · Journal of Sports Sciences · doi:10.1080/02640414.2020.1841394 · Cited by 13 · see citing works
VIT and Bi-LSTM for Micro-Expressions Recognition
Chen, H., Cui, J., Zhang, Y., & Zhang, Y. (2022) · IEEE International Conference on Information Systems and Computer Aided Education (ICISCAE) · doi:10.1109/iciscae55891.2022.9927522 · Cited by 7 · see citing works
Author Correction: Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality
Yuan, H., Plekhanova, T., Walmsley, R., Reynolds, A. C., Maddison, K. J., Bućan, M., Gehrman, P., Rowlands, A. V., Ray, D. W., Bennett, D., et al. (2024) · npj Digital Medicine · doi:10.1038/s41746-024-01148-y · Cited by 2 · see citing works

Flagged first, then by citations

Sorted by citations

A hybrid deep learning model based on vision transformer and convolutional neural networks for land use and land cover classification
Mahendra, H. N., Pushpalatha, V., Mallikarjunaswamy, S., & Subramoniam, S. R. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.114775 · Cited by 4 · see citing works
ViT-BiLSTM Multimodal Learning for Paediatric ADHD Recognition: Integrating Wearable Sensor Data with Clinical Profiles
Wang, L., & Yang, G. (2025) · Sensors · doi:10.3390/s25206459 · Cited by 3 · see citing works
Self-citation
WiVi-UF: Unified feature learning in cross-modal transformers with WiFi and vision data fusion for enhanced human activity recognition
Lin, X., Zhu, X., & Cambria, E. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.115548 · Cited by 3 · see citing works
Machine and Deep Learning for Detection of Moderate-to-Vigorous Physical Activity From Accelerometer Data: Systematic Scoping Review
Zi, Y., van de Ven, S. R. B., de Geus, E. J. C., & Chen, P. (2025) · Interactive Journal of Medical Research · doi:10.2196/76601 · Cited by 3 · see citing works
XNet: Enhancing Physical Activity Intensity Assessment With Attentional Multidomain Fusion and Visual Analytics
Mendoza, A., da Silva, E. P., Vega‐Oliveros, D. A., de Souza, T. F., Soriano-Vargas, A., Uchida, M., & Rocha, A. (2026) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2026.3666726 · Cited by 1 · see citing works
Towards Prediabetes Detection from Generative AI-Augmented Continuous Glucose Monitoring of Postprandial Glycemic Response using Bi-LSTM
Piersanti, A., Giudice, L. L. D., Marraudino, F., Göbl, C. S., Burattini, L., Tura, A., & Morettini, M. (2025) · IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE) · doi:10.1109/metroxraine66377.2025.11340102
Efficient Gastric Cancer Categorization Using Hybrid Deep Feature Extraction and Multi-Head Self-Attention
Wasima, J., Cruze, F. R. D., Hosen, M. F., Ali, M. S., & Kader, M. N. (2026) · IEEE International Conference on Quantum Photonics, Artificial Intelligence & Networking (QPAIN) · doi:10.1109/qpain69676.2026.11545560
Geometry-Aware 1D Residual CNN-BiLSTM Network for Automated Vertebral Localization and Scoliosis Classification
Kuo, P., & Chang, Y.-C. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8090286

Additional file 1 of A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration
Wang, L., Luo, Z., & Zhang, T. (2025) · Figshare · doi:10.6084/m9.figshare.28327699
Self-citation
Additional file 4 of A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration
Wang, L., Luo, Z., & Zhang, T. (2025) · Figshare · doi:10.6084/m9.figshare.28327708
Self-citation
Additional file 5 of A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration
Wang, L., Luo, Z., & Zhang, T. (2025) · Figshare · doi:10.6084/m9.figshare.28327711
Self-citation
Additional file 3 of A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration
Wang, L., Luo, Z., & Zhang, T. (2025) · Figshare · doi:10.6084/m9.figshare.28327705
Self-citation
Additional file 2 of A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration
Wang, L., Luo, Z., & Zhang, T. (2025) · Figshare · doi:10.6084/m9.figshare.28327702
Self-citation

Authors: Wang, L., Luo, Z., & Zhang, T.
Affiliations: University of Exeter (GB)
Yangtze University (CN)
University of Liverpool (GB)
Corresponding author: Wang, Lin
ORCID: Wang, Lin
Luo, Zizhang
Venue: BMC Biomedical Engineering
Publisher: Springer Science and Business Media LLC
Published: 2025-02-01
Volume: 7
Issue: 1
Pages: 2-2
ISSN: 2524-4426 (ISSN-L)
DOI: 10.1186/s42490-025-00088-2
PMC: PMC11786420
OpenAlex: W4407024896
Cited by: 8
References: 36
Citations by year: 2026: 5 · 2025: 3
FWCI: 2.81
Type: journal article
Language: en
Open access: gold
doi.org
Licence: cc-by
creativecommons.org/licenses/by/4.0 (tdm)
creativecommons.org/licenses/by/4.0 (vor)
Locations: BMC Biomedical Engineering (publishedVersion, cc-by)
Research Square (acceptedVersion, cc-by)
PubMed Central (publishedVersion, cc-by)
PubMed (publishedVersion)
DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders: China Scholarship Council
Topics: Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
Health Sciences › Medicine › Physiology › Physical Activity and Health
Physical Sciences › Computer Science › Signal Processing › Time Series Analysis and Forecasting
Keywords: physical activity intensity classification, vision transformer, bidirectional LSTM, hybrid deep learning model, accelerometer data, sedentary behavior, moderate-to-vigorous physical activity, light physical activity

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

A novel ViT-BILSTM model for physical activity intensity classification in adults using gravity-based acceleration

Abstract

Integrity signals

Claims the paper makes

Self-check

Cites

Cited by

Related

Record

About these checks