The Impact of YOLO Algorithms Within Fall Detection Application: A Review

Abstract

The phenomenon of human falls is a highly significant health concern, particularly for elderly individuals and disabled individuals who reside alone. The global elderly population is steadily growing. Consequently, human fall detection is emerging as a highly efficient method for enhancing the quality of life for individuals in need of assistance. Deep learning (DL) and computer vision have been extensively employed for assistive living purposes. This review article focuses on advanced DL model known as YOLO algorithm for non-intrusive fall detection using vision-based methods. Furthermore, we describe the evaluation architecture of YOLO algorithms (YOLOv1 - YOLOv8). Additionally, we provide a survey on benchmark datasets for fall detection. To enhance comprehension, we will provide a concise overview of various metrics employed to assess the efficacy of fall detection systems. This article also provides insight into future advancements in YOLO algorithms for detecting human falls.

Keywords

• Computer vision
• deep learning
• object detection
• YOLO
• real-time detection
• dataset