YOLOv7-Based Fall Detection for Enhanced Safety in Apartment Management Systems
Phuong Anh Nguyen, Duy Nguyen, Giang Nguyen, An Nguyen, Le Anh Ngoc · 2025
Apartment management systems play a crucial role in ensuring resident safety and well-being, particularly in large residential complexes. One key challenge is the detection of fall incidents, which are common among elderly residents and children. Effective fall detection, especially in common areas like hallways and lobbies, is critical to prompt emergency response and injury prevention. This paper introduces a novel approach utilizing YOLOv7, a deep learning-based object detection model, to monitor and detect falls in real-time using cameras strategically placed in these shared spaces. The proposed system aims to enhance coverage, improve privacy by avoiding the use of wearable devices, and optimize cost-efficiency. Evaluation results show a 95% detection accuracy, significantly reducing false alarms. This solution not only provides peace of mind for residents but also supports apartment managers in minimizing risks and enhancing overall safety.