Fall Detection Systems and Facial Recognition for Elderly Monitoring: A Systematic Review

Falls are a leading cause of injury-related hospitalizations in older adults, driving the need for automated fall detection systems (FDS) for timely intervention. This systematic review synthesizes recent advances in fall detection across wearable, vision-based, and hybrid approaches, with an emphasis on deep learning-based video systems. Guided by Kitchenham’s methodology, publications from 2015 to 2025 were searched across Scopus, IEEE Xplore, Google Scholar, and SciELO, yielding 79 primary studies from 778 initial articles (14 on elderly falls, 40 on fall detection systems, and 25 on facial recognition). The review addresses four research questions covering the causes and consequences of falls, the current state of FDS, their support for user-specific adaptation, and the feasibility of integrating facial recognition for personalized response. Key findings revealed that two-stage pose-based methods achieved the highest accuracy in controlled settings but degrade significantly under occlusion and multi-person scenarios, while single-stage detectors offer faster inference at the cost of reduced precision. A critical gap identified was the lack of identity-aware monitoring: existing FDS detect fall events but cannot identify which individual fell, preventing personalized emergency response in multi-occupancy settings. We explored facial recognition integration as a secondary identity layer, analyzing feasibility under fall-specific constraints including partial occlusions, low-resolution overhead capture, and real-time processing. Models trained with datasets such as VGGFace2 emerge as most suitable, while homomorphic encryption addresses privacy concerns in residential care. This work provides a comprehensive FDS evaluation, synthesized performance insights, and a framework for identity-aware monitoring that balances accuracy, efficiency, and privacy protection, supporting the development of personalized emergency response in elderly care environments.