Intelligent Fall Monitoring System for Senior Safety Using Image Processing

Abstract

With the growing challenge of an aging population and a shortage of caregivers, efficient fall detection systems are becoming essential in caregiving environments. This study presents a deep learning-based solution using a Sequential Convolutional Neural Network (CNN) model combined with drone mobility and advanced image recognition techniques. The system tracks care recipients in real-time and accurately identifies fall directions—forward, backward, leftward, and rightward. By enhancing Open Pose’s multi-person action analysis capabilities, the proposed model achieves an overall accuracy of 85% in detecting falls across diverse scenarios. Experimental results demonstrate significant improvement over baseline methods, with forward and leftward fall detection showing notable enhancements. These findings highlight the potential of integrating CNNs with mobility and action analysis technologies to enable reliable fall detection, ensuring timely responses and improved safety in caregiving environments.