SODSR:一种基于优化组合的三阶段超分辨率小目标检测方法

IF 5.3 3区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Xiaoyong Mei;Kejin Zhang;Changqin Huang;Xiao Chen;Ming Li;Zhao Li;Weiping Ding;Xindong Wu
{"title":"SODSR:一种基于优化组合的三阶段超分辨率小目标检测方法","authors":"Xiaoyong Mei;Kejin Zhang;Changqin Huang;Xiao Chen;Ming Li;Zhao Li;Weiping Ding;Xindong Wu","doi":"10.1109/TETCI.2024.3452749","DOIUrl":null,"url":null,"abstract":"Face detection is a fundamental task in computer vision, yet remains challenging in educational settings due to the presence of objects of various sizes. Subpar detection can significantly impede subsequent tasks' performance. To address this, we present a novel framework, Small Object Detection Super Resolution (SODSR), inspired by super resolution (SR) techniques for feature-level images. SODSR comprises three stages: (1) Constructing a 3D model evaluation matrix to select optimal model combinations based on detection accuracy and image quality metrics. (2) Employing Double-thread FDN in the first stage to preprocess images, enhancing feature resolution for potential facial objects. (3) Leveraging Multi-head HyperNet in the second stage to augment face feature detection and improve accuracy. Finally, in the third stage, we introduce AFPGAN, a facial prior feature enhancement network, coupled with StyleGAN2 for texture and contour detail enhancement. Experimental results demonstrate that SODSR outperforms existing small object detection (SOD) models in both accuracy and visual fidelity.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":"9 3","pages":"2410-2426"},"PeriodicalIF":5.3000,"publicationDate":"2024-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"SODSR: A Three-Stage Small Object Detection via Super-Resolution Using Optimizing Combination\",\"authors\":\"Xiaoyong Mei;Kejin Zhang;Changqin Huang;Xiao Chen;Ming Li;Zhao Li;Weiping Ding;Xindong Wu\",\"doi\":\"10.1109/TETCI.2024.3452749\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Face detection is a fundamental task in computer vision, yet remains challenging in educational settings due to the presence of objects of various sizes. Subpar detection can significantly impede subsequent tasks' performance. To address this, we present a novel framework, Small Object Detection Super Resolution (SODSR), inspired by super resolution (SR) techniques for feature-level images. SODSR comprises three stages: (1) Constructing a 3D model evaluation matrix to select optimal model combinations based on detection accuracy and image quality metrics. (2) Employing Double-thread FDN in the first stage to preprocess images, enhancing feature resolution for potential facial objects. (3) Leveraging Multi-head HyperNet in the second stage to augment face feature detection and improve accuracy. Finally, in the third stage, we introduce AFPGAN, a facial prior feature enhancement network, coupled with StyleGAN2 for texture and contour detail enhancement. Experimental results demonstrate that SODSR outperforms existing small object detection (SOD) models in both accuracy and visual fidelity.\",\"PeriodicalId\":13135,\"journal\":{\"name\":\"IEEE Transactions on Emerging Topics in Computational Intelligence\",\"volume\":\"9 3\",\"pages\":\"2410-2426\"},\"PeriodicalIF\":5.3000,\"publicationDate\":\"2024-09-20\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE Transactions on Emerging Topics in Computational Intelligence\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://ieeexplore.ieee.org/document/10684780/\",\"RegionNum\":3,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Emerging Topics in Computational Intelligence","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10684780/","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

摘要

人脸检测是计算机视觉的一项基本任务,但由于存在各种大小的物体,在教育环境中仍然具有挑战性。次优检测会显著影响后续任务的表现。为了解决这个问题,我们提出了一个新的框架,小目标检测超分辨率(SODSR),灵感来自特征级图像的超分辨率(SR)技术。SODSR包括三个阶段:(1)构建三维模型评价矩阵,根据检测精度和图像质量指标选择最优模型组合。(2)第一阶段采用双线FDN对图像进行预处理,增强对潜在面部目标的特征分辨率。(3)第二阶段利用多头HyperNet增强人脸特征检测,提高准确率。最后,在第三阶段,我们引入了人脸先验特征增强网络AFPGAN,并结合StyleGAN2进行纹理和轮廓细节增强。实验结果表明,SODSR在精度和视觉保真度方面都优于现有的小目标检测(SOD)模型。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
SODSR: A Three-Stage Small Object Detection via Super-Resolution Using Optimizing Combination
Face detection is a fundamental task in computer vision, yet remains challenging in educational settings due to the presence of objects of various sizes. Subpar detection can significantly impede subsequent tasks' performance. To address this, we present a novel framework, Small Object Detection Super Resolution (SODSR), inspired by super resolution (SR) techniques for feature-level images. SODSR comprises three stages: (1) Constructing a 3D model evaluation matrix to select optimal model combinations based on detection accuracy and image quality metrics. (2) Employing Double-thread FDN in the first stage to preprocess images, enhancing feature resolution for potential facial objects. (3) Leveraging Multi-head HyperNet in the second stage to augment face feature detection and improve accuracy. Finally, in the third stage, we introduce AFPGAN, a facial prior feature enhancement network, coupled with StyleGAN2 for texture and contour detail enhancement. Experimental results demonstrate that SODSR outperforms existing small object detection (SOD) models in both accuracy and visual fidelity.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
CiteScore
10.30
自引率
7.50%
发文量
147
期刊介绍: The IEEE Transactions on Emerging Topics in Computational Intelligence (TETCI) publishes original articles on emerging aspects of computational intelligence, including theory, applications, and surveys. TETCI is an electronics only publication. TETCI publishes six issues per year. Authors are encouraged to submit manuscripts in any emerging topic in computational intelligence, especially nature-inspired computing topics not covered by other IEEE Computational Intelligence Society journals. A few such illustrative examples are glial cell networks, computational neuroscience, Brain Computer Interface, ambient intelligence, non-fuzzy computing with words, artificial life, cultural learning, artificial endocrine networks, social reasoning, artificial hormone networks, computational intelligence for the IoT and Smart-X technologies.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信