A COMPREHENSIVE REVIEW OF OBJECT DETECTION IN YOLO:EVOLUTION, VARIANTS, AND APPLICATIONS

Abstract

In computer vision, object detection is a crucial task with
applications ranging from autonomous systems to surveillance. In this field, the You Only Look Once (YOLO) algorithm has become a major player, providing real-time detection with its unified architecture that manages object localization and categorization at the same time. This essay examines the development of YOLO, covering its early iterations as well as more recent developments like
YOLOv4. It looks at the algorithm's uses in a variety of industries, including driverless vehicles and imaging in medicine. Notwithstanding its achievements, problems still exist, and new approaches are being suggested by continuing study. The most recent additions to YOLOv4, such as CSPNet and PANet,
represent a major improvement in scalability and performance. The impact of YOLO on object detection is highlighted in the paper's conclusion, along with its adaptability and the ongoing study that is required to solve problems and improve capabilities.