Abstract: Falls among individuals pose serious risks to their health and independence, underscoring the importance of effective fall detection solutions. This study aims to address this critical issue by proposing a novel approach that integrates Computer Vision and Deep Learning for Real-time fall detection and assistance.
Traditionally, fall detection systems have relied on wearable sensors, which, despite their widespread use, often suffer from drawbacks such as false alarms and discomfort for the wearer. In response to these limitations, this project introduces an efficient solution by leveraging Computer Vision and Deep Learning.
The core of this innovative system lies in the integration of the YOLOv8 (You Only Look Once) which is a cutting-edge, real-time object detection algorithm that uses Convolutional Neural Network (CNN) to predict the bounding boxes and class probabilities of objects in input images with Computer Vision. YOLOv8, a variant of the YOLO object algorithm series, has demonstrated superior performance in identifying various objects, and therefore has been used in detecting fall events, with remarkable accuracy and efficiency.
By combining the strengths of YOLOv8 and Computer Vision, this solution offers improved accuracy and reliability in identifying fall events and also enhances the overall user experience by providing timely assistance and ensuring the safety and well-being of individuals.
Keywords: Computer Vision, Deep Learning, Microcontrollers, bounding boxes
|
DOI:
10.17148/IJARCCE.2024.13549
