A Real-time Fall Detection Model based on Body Landmark Tracking for Intelligent Healthcare

An unintentional fall, which can occasionally even result in death, is one of the main concerns for the loved ones of the elderly. As shown by recent research, vision-based methods enable the real-time surveillance of such occurrences, enabling worried caregivers or loved ones to assist elderly people as soon as possible. The design and implementation of a real-time vision-based fall detection method utilizing the mediapipe framework is our primary objective in this article. With the aid of computer vision techniques for identifying body landmarks, the model can thoroughly examine variations in movement patterns and posture angles. Experimental results demonstrate the efficacy of the proposed model in accurately identifying fall events while minimizing errors in detecting ADL. The proposed approach outperforms existing techniques in the area of fall detection among senior adults, achieving an accuracy score of 98.66%.