A Comprehensive Review of YOLOv8-MediaPipe and Deep Learning-Based Multi-Person Fall Detection System

This review has synthesized modern deep learning architectures for multi-person tracking and real-time fall detection. The literature shows a clear paradigm shift away from traditional geometric heuristics and toward top-down hybrid frameworks. By pairing object detectors like YOLOv8 with robust tracking mechanisms (ByteTrack) and isolated pose estimators (Me-diaPipe), contemporary systems successfully mitigate historical issues like tracking loss and identity switching in crowded settings. When combined with sequence modelers like LSTMs, these systems provide the low latency and reliability needed for deployment in smart healthcare facilities, elderly care homes, and autonomous safety networks. Future advancements will de-pend on optimizing these multi-stage models for edge computing and ensuring robustness across unpredictable real-world environments.  

[1]       T. Liu and S. Lee, “A review of fall detection systems based on deep learning,” IEEE Access, vol. 7, pp. 142409–142421, 2019. [2]       M. Mubashir, L. Shao, and L. Seed, “A survey on fall detection: Principles and approaches,” Neurocomputing, vol. 100, pp. 144–152, 2013. [3]       M. Yu et al., “A survey on vision-based fall detection,” Biomedical Signal Processing and Control, vol. 4, no. 1, pp. 45–56, 2009. [4]       World Health Organization, “Falls fact sheet,” 2021. [5]       E. Alam et al., “GMDCSA-24: A dataset for human fall detection in videos,” Data in Brief, vol. 57, 2024. [6]       W. Liao, T. Chen, and J. Yang, “A novel fall detection system using YOLOV3 and pose estimation,” Proc. IEEE ICIP, pp. 3704–3708, 2020. [7]       S. Chien and T. Li, “A real-time fall detection system using pose estimation and machine learning,” Journal of Healthcare Engineering, vol. 2019, pp. 1–9, 2019. [8]       J. Redmon et al., “You Only Look Once: Unified, real-time object detection,” Proc. IEEE CVPR, pp. 779–788, 2016. [9]       A. Sixsmith and N. Johnson, “A smart sensor to detect the falls of the elderly,” IEEE Pervasive Computing, vol. 3, no. 2, pp. 42–47, 2004. [10]    N. Thome, S. Miguet, and S. Ambellouis, “A real-time, multiview fall detection system,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 18, no. 11, pp. 1522–1527, 2008.