摘要
概述目标检测算法发展历程的基础上,重点分析和综述了基于关键点的目标检测算法CornerNet、ExtremeNet、CornerNet-Lite、两种CenterNet和FCOS,并与基于锚点的目标检测框架进行性能对比,总结提出基于关键点的目标检测框架是未来热门研究方向。
This paper briefly summarizes the development of the boject detection algorithm,focuses on the analysis and overview of the object detection algorithms based on key points,such as CornerNet,ExtremeNet,CornerNet-Lite,two kinds of CenterNet and FCOS and performance comparison,compared with the anchor based target detection framework,and summarizes the key target detection framework is a hot research direction in the future.
出处
《信息技术与标准化》
2020年第6期44-48,52,共6页
Information Technology & Standardization
关键词
目标检测
深度学习
关键点
算法
object detection
deep learning
key points
algorithms
