机床与液压2024,Vol.52Issue(1) :202-209.DOI:10.3969/j.issn.1001-3881.2024.01.031

基于深度学习的无锚框目标检测算法综述

Overview of Anchor-Free Object Detection Algorithms Based on Depth Learning

高海涛 朱超涵 张天棋 郝飞 茅新宇
机床与液压2024,Vol.52Issue(1) :202-209.DOI:10.3969/j.issn.1001-3881.2024.01.031

基于深度学习的无锚框目标检测算法综述

Overview of Anchor-Free Object Detection Algorithms Based on Depth Learning

高海涛 1朱超涵 1张天棋 1郝飞 1茅新宇1
扫码查看

作者信息

  • 1. 南京工程学院机械工程学院,江苏南京 211167
  • 折叠

摘要

近年来,基于深度学习的无锚框目标检测算法备受关注.为了深入理解无锚框检测算法,对比分析了基于深度学习的无锚框检测算法的原理机制、网络结构、核心特性以及优缺点,归纳总结了无锚框检测算法的核心技术,并在同一数据集上通过性能实验研究上述算法的性能,总结提出基于深度学习的目标检测算法未来的研究方向.

Abstract

In recent years,target detection algorithm based on deep learning has attracted much attention.In order to deeply under-stand the typical anchor-free object detection algorithms,the principle mechanism,network structure,core characteristics,advantages and disadvantages of seven anchor-free object detection algorithms based on deep learning were compared and analyzed.The perform-ance of the above algorithms was experimentally studied.On this basis,the main characteristics of the anchor-free object detection algo-rithm were summarized,and the research direction of the anchor-free object detection algorithm was pointed out.

关键词

无锚框目标检测算法/深度学习/算法比较

Key words

anchor-free object detection algorithm/deep learning/algorithm comparison

引用本文复制引用

基金项目

国家自然科学基金青年科学基金(51705238)

江苏省研究生实践创新计划项目(SJCX_0916)

江苏省研究生实践创新计划项目(SJCX23_1173)

江苏省现代农机装备与技术示范推广项目(NJ2021-58)

出版年

2024
机床与液压
中国机械工程学会 广州机械科学研究院有限公司

机床与液压

CSTPCD北大核心
影响因子:0.32
ISSN:1001-3881
参考文献量34
段落导航相关论文