{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T17:07:37Z","timestamp":1791911257159,"version":"4.3.4"},"reference-count":42,"publisher":"MDPI AG","issue":"21","license":[{"start":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T00:00:00Z","timestamp":1762387200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Fall detection systems are a critical application of human pose estimation, frequently struggle with achieving real-world robustness due to their reliance on domain-specific datasets and a limited capacity for generalization to novel conditions. Models trained on controlled, canonical camera views often fail when subjects are viewed from new perspectives or are partially occluded, resulting in missed detections or false positives. This study tackles these limitations by proposing the Viewpoint Invariant Robust Aggregation Graph Convolutional Network (VIRA-GCN), an adaptation of the Richly Activated GCN for fall detection. The VIRA-GCN introduces a novel dual-strategy solution: a synthetic viewpoint generation process to augment training data and an efficient inference-time aggregation method to form consensus-based predictions. We demonstrate that augmenting the Le2i dataset with simulated rotations and occlusions allows a standard pose estimation model to achieve a significant increase in its fall detection capabilities. The VIRA-GCN achieved 99.81% accuracy on the Le2i dataset, confirming its enhanced robustness. Furthermore, the model is suitable for low-resource deployment, utilizing only 4.06 M parameters and achieving a real-time inference latency of 7.50 ms. This work presents a practical and efficient solution for developing a single-camera fall detection system robust to viewpoint variations, and introduces a reusable mapping function to convert Kinect data to the MMPose format, ensuring consistent comparison with state-of-the-art models.<\/jats:p>","DOI":"10.3390\/s25216783","type":"journal-article","created":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T09:02:03Z","timestamp":1762419723000},"page":"6783","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Robust 3D Skeletal Joint Fall Detection in Occluded and Rotated Views Using Data Augmentation and Inference\u2013Time Aggregation"],"prefix":"10.3390","volume":"25","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-9803-4729","authenticated-orcid":false,"given":"Maryem","family":"Zobi","sequence":"first","affiliation":[{"name":"ADMIR Laboratory, National School of Computer Science and Systems Analysis, Mohammed V University, Rabat 10000, Morocco"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-0475-5891","authenticated-orcid":false,"given":"Lorenzo","family":"Bolzani","sequence":"additional","affiliation":[{"name":"flAIght, Viale dell\u2019Innovazione 13, 20126 Milan, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7030-3910","authenticated-orcid":false,"given":"Youness","family":"Tabii","sequence":"additional","affiliation":[{"name":"ADMIR Laboratory, National School of Computer Science and Systems Analysis, Mohammed V University, Rabat 10000, Morocco"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9736-7260","authenticated-orcid":false,"given":"Rachid","family":"Oulad Haj Thami","sequence":"additional","affiliation":[{"name":"ADMIR Laboratory, National School of Computer Science and Systems Analysis, Mohammed V University, Rabat 10000, Morocco"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2025,11,6]]},"reference":[{"key":"ref_1","unstructured":"The United Nations Department of Economic and Social Affairs (UN DESA) (2020). World Population Ageing 2020 Highlights: Living Arrangements of Older Persons, United Nations. (ST\/ESA\/SER.A\/451)."},{"key":"ref_2","unstructured":"WHO (2021). WHO Global Report on Falls Prevention in Older Age."},{"key":"ref_3","first-page":"808","article-title":"A review of vision-based fall detection systems for elderly people: A performance analysis","volume":"32","author":"Nazar","year":"2020","journal-title":"J. King Saud-Univ. Comput. Inf. Sci."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"94249","DOI":"10.1109\/ACCESS.2022.3203174","article-title":"Exploring Human Pose Estimation and the Usage of Synthetic Data for Elderly Fall Detection in Real-World Surveillance","volume":"10","author":"Shaha","year":"2022","journal-title":"IEEE Access"},{"key":"ref_5","first-page":"2390","article-title":"Robust Self-Adaptation Fall-Detection System Based on Camera Height","volume":"19","author":"Zhang","year":"2019","journal-title":"Sensors"},{"key":"ref_6","first-page":"262","article-title":"Fall Detection with Multiple Cameras: An Occlusion-Resistant Method Based on 3-D Silhouette Vertical Distribution","volume":"24","author":"Sannino","year":"2020","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"23","DOI":"10.1109\/TCE.2014.6780921","article-title":"An enhanced fall detection system for elderly person monitoring using consumer home networks","volume":"60","author":"Wang","year":"2014","journal-title":"IEEE Trans. Consum. Electron."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Desai, K., Mane, P., Dsilva, M., Zare, A., Shingala, P., and Ambawade, D. (2020, January 2\u20134). A novel machine learning based wearable belt for fall detection. Proceedings of the 2020 IEEE International Conference on Computing, Power and Communication Technologies (GUCON), Online.","DOI":"10.1109\/GUCON48875.2020.9231114"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"38670","DOI":"10.1109\/ACCESS.2019.2906693","article-title":"A Machine Learning Approach for Fall Detection and Daily Living Activity Recognition","volume":"7","author":"Chelli","year":"2019","journal-title":"IEEE Access"},{"key":"ref_10","first-page":"150","article-title":"OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields","volume":"43","author":"Cao","year":"2019","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_11","unstructured":"MMPose Contributors (2025, October 19). MMPose: An OpenMMLab Pose Estimation Toolbox and Benchmark. Available online: https:\/\/github.com\/open-mmlab\/mmpose."},{"key":"ref_12","first-page":"1363","article-title":"An improved feature-based method for fall detection","volume":"26","author":"Yao","year":"2019","journal-title":"Teh. Vjesn."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"430","DOI":"10.1109\/JBHI.2014.2319372","article-title":"Fall Detection Based on Body Part Tracking Using a Depth Camera","volume":"19","author":"Bian","year":"2015","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Debard, G., Karsmakers, P., Deschodt, M., Vlaeyen, E., Bergh, J., Dejaeger, E., Milisen, K., Goedem\u00e9, T., Tuytelaars, T., and Vanrumste, B. (2011, January 25\u201328). Camera Based Fall Detection Using Multiple Features Validated with Real Life Video. Proceedings of the 7th International Conference on Intelligent Environments, Nottingham, UK.","DOI":"10.3233\/978-1-60750-795-6-441"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"314","DOI":"10.1109\/JBHI.2018.2808281","article-title":"Deep Learning for Fall Detection: Three-Dimensional CNN Combined with LSTM on Video Kinematic Data","volume":"23","author":"Lu","year":"2019","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Tian, Y., Lee, G.-H., He, H., Hsu, C.-Y., and Katabi, D. (2018, January 8\u201312). Rf-based fall monitoring using convolutional neural networks. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, Singapore.","DOI":"10.1145\/3264947"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"161337","DOI":"10.1109\/ACCESS.2020.3021795","article-title":"Vision-based fall event detection in complex background using attention guided bi-directional lstm","volume":"8","author":"Chen","year":"2020","journal-title":"IEEE Access"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Dutt, M., Gupta, A., Goodwin, M., and Omlin, C.W. (2024). An interpretable modular deep learning framework for video-based fall detection. Appl. Sci., 14.","DOI":"10.3390\/app14114722"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"153049","DOI":"10.1109\/ACCESS.2019.2947518","article-title":"Implementation of fall detection system based on 3D skeleton for deep learning technique","volume":"7","author":"Tsai","year":"2019","journal-title":"IEEE Access"},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"28224","DOI":"10.1109\/ACCESS.2021.3058219","article-title":"Vision-Based Fall Detection Using ST-GCN","volume":"9","author":"Keskes","year":"2021","journal-title":"IEEE Access"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Egawa, R., Miah, A.S.M., Hirooka, K., Tomioka, Y., and Shin, J. (2023). Dynamic fall detection using graph-based spatial temporal convolution and attention network. Electronics, 12.","DOI":"10.3390\/electronics12153234"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Shin, J., Miah, A.S.M., Egawa, R., Hirooka, K., Hasan, M.A.M., Tomioka, Y., and Hwang, Y.S. (2025). Fall recognition using a three stream spatio temporal GCN model with adaptive feature aggregation. Sci. Rep., 15.","DOI":"10.1038\/s41598-025-95508-7"},{"key":"ref_23","unstructured":"Kepski, M., and Kwolek, B. (2014). A New Multimodal Approach to Fall Detection. PLoS ONE, 9."},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Shahroudy, A., Liu, J., Ng, T.T., and Wang, G. (2016, January 27\u201330). NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.","DOI":"10.1109\/CVPR.2016.115"},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"043009","DOI":"10.1117\/1.JEI.22.4.041106","article-title":"Optimised spatio-temporal descriptors for real-time fall detection: Comparison of SVM and Adaboost based classification","volume":"22","author":"Charfi","year":"2013","journal-title":"J. Electron. Imaging"},{"key":"ref_26","unstructured":"Kepski, M., and Kwolek, B. (2019, January 16\u201318). A New Fall Detection Dataset and the Comparison of Fall Detection Methods. Proceedings of the 2019 International Conference on Computer Vision and Graphics, Warsaw, Poland."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"1915","DOI":"10.1109\/JBHI.2014.2304357","article-title":"Depth-based human fall detection via shape features and improved extreme learning machine","volume":"18","author":"Ma","year":"2014","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Soto-Valero, L.G., and Medina-Merodio, M. (2017). SisFall: A Fall and Movement Dataset. Sensors, 17.","DOI":"10.3390\/s17010198"},{"key":"ref_29","unstructured":"(2025, October 19). TST Fall Detection Dataset v2. IEEE Dataport, 2016. Available online: https:\/\/ieee-dataport.org\/documents\/tst-fall-detection-dataset-v2."},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Yu, X., Jang, J., and Xiong, S. (2021). KFall: A Comprehensive Motion Dataset to Detect Pre-impact Fall for the Elderly Based on Wearable Inertial Sensors. Front. Aging Neurosci., 13.","DOI":"10.3389\/fnagi.2021.692865"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Vavoulas, G., Chatzaki, C., Malliotakis, T., and Pediaditis, M. (2016, January 21\u201322). The MobiAct Dataset: Recognition of Activities of Daily Living using Smartphones. Proceedings of the 2nd International Conference on Information and Communication Technologies for Ageing Well and e-Health, Rome, Italy.","DOI":"10.5220\/0005792401430151"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"5110","DOI":"10.1109\/JSEN.2020.3018335","article-title":"FallAllD: An Open Dataset of Human Falls and Activities of Daily Living for Classical and Deep Learning Applications","volume":"21","author":"Saleh","year":"2021","journal-title":"IEEE Sens. J."},{"key":"ref_33","first-page":"1339","article-title":"The FARSEEING real-world fall repository: A large-scale collaborative database to collect and share sensor signals from real-world falls","volume":"63","author":"Klenk","year":"2016","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"7157","DOI":"10.1109\/TPAMI.2022.3222784","article-title":"AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time","volume":"45","author":"Fang","year":"2022","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_35","unstructured":"Lugaresi, C., Tang, J., Nash, H., McClanahan, C., Uboweja, E., Hays, M., Zhang, F., Chang, C.-L., Yong, M.G., and Lee, J. (2019). MediaPipe: A Framework for Building Perception Pipelines. arXiv."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"1915","DOI":"10.1109\/TCSVT.2020.3015051","article-title":"Richly Activated Graph Convolutional Network for Robust Skeleton-based Action Recognition","volume":"31","author":"Song","year":"2021","journal-title":"IEEE Trans. Circuits Syst. Video Technol."},{"key":"ref_37","unstructured":"Li, W., Wang, Y., Wang, Y., Li, B., Zhang, H., and Wu, Y. (2021, January 6\u201312). Lightweight Spatio-Temporal Graph Convolutional Networks for Skeleton-based Action Recognition. Proceedings of the 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Virtual."},{"key":"ref_38","unstructured":"Alaoui, A., El-Yaagoubi, M., and Chehri, A. (2021, January 19\u201320). V2V-PoseNet: A View-Invariant Pose Estimation Network for Action Recognition. Proceedings of the International Conference on Advanced Communication Technologies and Information Systems (ICACTIS), Coimbatore, India."},{"key":"ref_39","first-page":"789","article-title":"YOLOv7-DeepSORT: An Improved Human Fall Detection Method","volume":"23","author":"Zi","year":"2023","journal-title":"Sensors"},{"key":"ref_40","first-page":"5565","article-title":"A Novel Fall Detection Method Based on AlphaPose and Spatio-Temporal Graph Convolutional Networks","volume":"23","author":"Li","year":"2023","journal-title":"Sensors"},{"key":"ref_41","first-page":"15993","article-title":"3D Multi-Stream CNNs for Real-Time Fall Detection","volume":"11","author":"Alanazi","year":"2023","journal-title":"IEEE Access"},{"key":"ref_42","unstructured":"Shin, J., Kim, S., and Kim, D. (2024, January 16\u201318). Three-Stream Spatio-Temporal Graph Convolutional Networks for Human Fall Detection. Proceedings of the 2024 International Conference on Information and Communication Technology Convergence (ICTC), Jeju Island, Republic of Korea."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/25\/21\/6783\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T09:05:44Z","timestamp":1762419944000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/25\/21\/6783"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,6]]},"references-count":42,"journal-issue":{"issue":"21","published-online":{"date-parts":[[2025,11]]}},"alternative-id":["s25216783"],"URL":"https:\/\/doi.org\/10.3390\/s25216783","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,6]]}}}