Overview of YOLO Object Detection Algorithm

Abstract

As an important research direction in the field of computer vision, object detection has developed 
 rapidly and many kinds of mature algorithms emerged. The series of YOLO (You Only Look Once) 
 algorithms implement one-stage detection based on regression ideas, which showing preeminent 
 in speed and owning strong generalization on a variety of datasets. This paper will give a simple 
 introduction to the current mainstream deep learning object detection algorithm, then focus on 
 combing the principle and optimizational process of the series of YOLO algorithms, summarize 
 the latest 
 breakthroughs in YOLO algorithm, Hopefully that can provide reference for the 
 research of related topics.