Comparison on Object Detection Algorithms: A Taxonomy

Visual object detection is a popular task, which categorizes all the defined objects in the whole images. With the emerging of numerous object detection frameworks, many detection methods have been proposed. In this paper, we aim to survey several distinguished detection methods, which consists of two-stage based methods and one-stage based methods. In specific, we summarize 9 different object detection papers, including RCNN based methods, YOLO, M2Det and CornerNet, etc. All these papers play important roles and obtain state-of-the-art performances. We summarize all the frameworks, motivations and training steps for all methods. We hope that our review can give an in-depth guidance for the beginners for recognizing this search fields.