Bridging the Appearance Domain Gap in Elderly Posture Recognition with YOLOv9
Source abstract
Abstract
Accurate posture detection of elderly people is crucial to improve monitoring and provide timely alerts in homes and elderly care facilities. Human posture recognition is experiencing a great leap in performance with the incorporation of deep neural networks (DNNs) such as YOLOv9. Unfortunately, DNNs require large amounts of annotated data for training, which can be addressed by using virtual reality images. This paper investigates how to address the appearance domain that lies between synthetic and natural images. Therefore, four experiments (VIRTUAL–VIRTUAL; HYBRID–VIRTUAL; VIRTUAL–REAL; and HYBRID–REAL) were designed to assess the feasibility of recognising the postures of virtual or real elderly people after training with virtual and real images of elderly people. The results show that YOLOv9 achieves the most outstanding accuracy of 98.41% in detecting and discriminating between standing, sitting, and lying postures after training on a large number of virtual images complemented by a much smaller number of real images when testing on real images.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Face recognition and analysis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Instituto de Investigación en Informática de Albacete, 02071 Albacete, Spain
Departamento de Ingeniería Eléctrica, Electrónica, Automática y Comunicaciones, Universidad de Castilla-La Mancha, 02071 Albacete, Spain
Computer Science and Communications Research Centre, Polytechnic Institute of Leiria, School of Technology and Management, 2411-901 Leiria, Portugal
INOV INESC INOVAÇÃO, Institute of New Technologies—Leiria Office, 2411-901 Leiria, Portugal
Departamento de Sistemas Informáticos, Universidad de Castilla-La Mancha, 02071 Albacete, Spain
Biomedical Research Networking Centre in Mental Health (CIBERSAM), 28016 Madrid, Spain
Funding
Spanish MICIU/AEI/10.13039/501100011033 · CIBERSAM · Instituto de Salud Carlos III · MICIU · Universidad de Castilla-La Mancha · Portuguese Foundation for Science and Technology (FCT), I.P. · Junta de Comunidades de Castilla-La Mancha/ESF
References (48)
Show source-provided references
1. Tuba, M., Akashe, S., and Joshi, A. (2023). Fall Detection of Elderly Using YOLO. Proceedings of the ICT Systems and Sustainability, Springer.10.1007/978-981-19-5221-0 ↗
2. Raman, B., Murala, S., Chowdhury, A., Dhall, A., and Goyal, P. (2022). Elderly Patient Fall Detection Using Video Surveillance. Proceedings of the Computer Vision and Image Processing, Springer.10.1007/978-3-031-11349-9 ↗
3. A fuzzy model for human fall detection in infrared video10.3233/IFS-2012-0548 ↗
4. Ferrández Vicente, J.M., Val Calvo, M., and Adeli, H. (2024). Heart Attack Detection Using Body Posture and Facial Expression of Pain. Proceedings of the Artificial Intelligence for Neuroscience and Emotional Systems, Springer.10.1007/978-3-031-61140-7 ↗
5. Comparison of RGB-D and IMU-based gesture recognition for human-robot interaction in remanufacturing10.1007/s00170-021-08125-9 ↗
6. Wang, Y., Chi, Z., Liu, M., Li, G., and Ding, S. (2023). High-Performance Lightweight Fall Detection with an Improved YOLOv5s Algorithm. Machines, 11.10.3390/machines11080818 ↗
7. Ferrández Vicente, J.M., Álvarez-Sánchez, J.R., de la Paz López, F., and Adeli, H. (2022). Vision-Based Human Posture Detection from a Virtual Home-Care Unmanned Aerial Vehicle. Proceedings of the Bio-Inspired Systems and Applications: From Robotics to Ambient Intelligence, Springer.10.1007/978-3-031-06527-9 ↗
8. Ebadi, S.E., Jhang, Y.C., Zook, A., Dhakad, S., Crespi, A., Parisi, P., Borkman, S., Hogins, J., and Ganguly, S. (2022). PeopleSansPeople: A Synthetic Data Generator for Human-Centric Computer Vision. arXiv.
9. Skeleton-based human activity recognition using ConvLSTM and guided feature learning10.1007/s00500-021-06238-7 ↗
10. Skeleton joint trajectories based human activity recognition using deep RNN10.1007/s11042-023-15024-6 ↗
11. Combining CNN streams of dynamic image and depth data for action recognition10.1007/s00530-019-00645-5 ↗
12. Human activity recognition with smartphone-integrated sensors: A survey10.1016/j.eswa.2024.123143 ↗
13. An elderly health monitoring system based on biological and behavioral indicators in internet of things10.1007/s12652-020-02579-7 ↗
14. Design of Health and Elderly Care Intelligent Monitoring System Based on IoT Wireless Sensing and Data Mining
15. A Survey on Artificial Intelligence in Posture Recognition
16. Lee, M.F.R., Chen, Y.C., and Tsai, C.Y. (2022). Deep Learning-Based Human Body Posture Recognition and Tracking for Unmanned Aerial Vehicles. Processes, 10.10.3390/pr10112295 ↗
17. Ogundokun, R.O., Maskeliūnas, R., and Damaševičius, R. (2022). Human Posture Detection Using Image Augmentation and Hyperparameter-Optimized Transfer Learning Algorithms. Appl. Sci., 12.10.3390/app121910156 ↗
18. Kviatkovsky, I., Bhonker, N., and Medioni, G. (2020). From Real to Synthetic and Back: Synthesizing Training Data for Multi-Person Scene Understanding. arXiv.
19. Deep unsupervised domain adaptation: A review of recent advances and perspectives10.1561/116.00000192 ↗
20. Video surveillance using deep transfer learning and deep domain adaptation: Towards better generalization10.1016/j.engappai.2022.105698 ↗
21. Domain adaptation: Challenges, methods, datasets, and applications10.1109/ACCESS.2023.3237025 ↗
22. Anvari, T., Park, K., and Kim, G. (2023). Upper Body Pose Estimation Using Deep Learning for a Virtual Reality Avatar. Appl. Sci., 13.10.3390/app13042460 ↗
23. Romero, A., Carvalho, P., Côrte-Real, L., and Pereira, A. (2023). Synthesizing Human Activity for Data Generation. J. Imaging, 9.10.3390/jimaging9100204 ↗
24. Reddy, A.V., Shah, K., Paul, W., Mocharla, R., Hoffman, J., Katyal, K.D., Manocha, D., de Melo, C.M., and Chellappa, R. (June, January 29). Synthetic-to-Real Domain Adaptation for Action Recognition: A Dataset and Baseline Performances. Proceedings of the 2023 IEEE International Conference on Robotics and Automation, London, UK.10.1109/ICRA48891.2023.10160416 ↗
25. Synthetic-real image domain adaptation for indoor camera pose regression using a 3D model10.1016/j.isprsjprs.2023.06.013 ↗
26. A Study of Using Synthetic Data for Effective Association Knowledge Learning10.1007/s11633-022-1380-x ↗
27. Yue, X., Zhang, Y., Zhao, S., Sangiovanni-Vincentelli, A., Keutzer, K., and Gong, B. (November, January 27). Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization without Accessing Target Domain Data. Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision, Seoul, Republic of Korea.10.1109/ICCV.2019.00219 ↗
28. Bustamante, A., Belmonte, L.M., Morales, R., Pereira, A., and Fernández-Caballero, A. (2022). Video Processing from a Virtual Unmanned Aerial Vehicle: Comparing Two Approaches to Using OpenCV in Unity. Appl. Sci., 12.10.3390/app12125958 ↗
29. Hennicke, L., Adriano, C.M., Giese, H., Koehler, J.M., and Schott, L. (2024). Mind the Gap Between Synthetic and Real: Utilizing Transfer Learning to Probe the Boundaries of Stable Diffusion Generated Data. arXiv.
30. Li, Y., Dong, X., Chen, C., Li, J., Wen, Y., Spranger, M., and Lyu, L. (2024). Is Synthetic Image Useful for Transfer Learning? An Investigation into Data Generation, Volume, and Utilization. arXiv.
31. Nowruzi, F.E., Kapoor, P., Kolhatkar, D., Hassanat, F.A., Laganière, R., and Rebut, J. (2019). How much real data do we actually need: Analyzing object detection performance using synthetic and real data. arXiv.
32. Dao, N.N., Thinh, T.N., and Nguyen, N.T. (2023). Building an AI-Powered IoT App for Fall Detection Using Yolov8 Approach. Proceedings of the Intelligence of Things: Technologies and Applications, Springer.
33. Lu, K.L., and Chu, E.T.H. (2018). An Image-Based Fall Detection System for the Elderly. Appl. Sci., 8.10.3390/app8101995 ↗
34. A Smartphone-Enabled Fall Detection Framework for Elderly People in Connected Home Healthcare10.1109/MNET.001.1900100 ↗
35. Killer heuristic optimized convolution neural network-based fall detection with wearable IoT sensor devices10.1016/j.measurement.2020.108258 ↗
36. Tateno, S., Meng, F., Qian, R., and Hachiya, Y. (2020). Privacy-Preserved Fall Detection Method with Three-Dimensional Convolutional Neural Network Using Low-Resolution Infrared Array Sensor. Sensors, 20.10.3390/s20205957 ↗
37. Contactless Fall Detection Using Time-Frequency Analysis and Convolutional Neural Networks10.1109/TII.2021.3049342 ↗
38. Minvielle, L., and Audiffren, J. (2019). NurseNet: Monitoring Elderly Levels of Activity with a Piezoelectric Floor. Sensors, 19.10.3390/s19183851 ↗
39. Zerkouk, M., and Chikhaoui, B. (2020). Spatio-Temporal Abnormal Behavior Prediction in Elderly Persons Using Deep Learning Models. Sensors, 20.10.3390/s20082359 ↗
40. Sensor-based activity recognition of solitary elderly via stigmergy and two-layer framework10.1016/j.engappai.2020.103859 ↗
41. Chutimawattanakul, P., and Samanpiboon, P. (2022, January 24–27). Fall Detection for The Elderly using YOLOv4 and LSTM. Proceedings of the 2022 19th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), Prachuap Khiri Khan, Thailand.10.1109/ECTI-CON54298.2022.9795534 ↗
42. Elderly Fall Detection Based on Improved YOLOv5s Network10.1109/ACCESS.2022.3202293 ↗
43. Raza, A., Yousaf, M.H., and Velastin, S.A. (2022, January 7–10). Human Fall Detection using YOLO: A Real-Time and AI-on-the-Edge Perspective. Proceedings of the 2022 12th International Conference on Pattern Recognition Systems (ICPRS), Saint-Etienne, France.10.1109/ICPRS54038.2022.9854070 ↗
44. Dai, Y., and Liu, W. (2023). GL-YOLO-Lite: A Novel Lightweight Fallen Person Detection Model. Entropy, 25.10.3390/e25040587 ↗
45. Li, Y., Wu, Y., Chen, X., Chen, H., Kong, D., Tang, H., and Li, S. (2023). Beyond Human Detection: A Benchmark for Detecting Common Human Posture. Sensors, 23.10.3390/s23198061 ↗
46. ARFDNet: An efficient activity recognition & fall detection system using latent feature pooling10.1016/j.knosys.2021.107948 ↗
47. Guerra, B.M.V., Ramat, S., Beltrami, G., and Schmid, M. (2023). Recurrent Network Solutions for Human Posture Recognition Based on Kinect Skeletal Data. Sensors, 23.10.3390/s23115260 ↗
48. Multimodal vision-based human action recognition using deep learning: A review10.1007/s10462-024-10730-5 ↗