中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 视频内容安全团队
Multi-Scale Low-Discriminative Feature Reactivation for Weakly Supervised Object Localization

文献类型：期刊论文

作者Wang, Bo1,2; Yuan, Chunfeng3; Li, Bing3,4; Ding, Xinmiao5; Li, Zeya6; Wu, Ying7; Hu, Weiming3,8,9
刊名IEEE TRANSACTIONS ON IMAGE PROCESSING
出版日期2021
卷号30页码:6050-6065
关键词Location awareness Neurons Task analysis Pattern recognition Automation Search problems Proposals Weakly supervised object localization feature recalibration multi-scale class activation mapping
ISSN号1057-7149
DOI10.1109/TIP.2021.3091833
通讯作者Yuan, Chunfeng(cfyuan@nlpr.ia.ac.cn)
英文摘要For weakly supervised object localization (WSOL), how to avoid the network focusing only on some small discriminative parts is a main challenge needed to solve. The widely-used Class Activation Mapping (CAM) based paradigm usually employs Adversarial Learning (AL) strategy to search more object parts by constantly hiding discovered object features, but the adversarial process is difficult to control. In this paper, we propose a novel CAM-based framework with Multi-scale Low-Discriminative Feature Reactivation (mLDFR) for WSOL. The mLDFR framework reactivates the low-discriminative object parts via bottom-up continuous feature maps recalibration and multi-scale object category mapping. Compared with the AL-based methods, our method fully improves the localization power of the network without damaging the classification power and can perform multi-instance localization, which are hard to achieve under the AL-based framework. Moreover, the mLDFR framework is flexible, and can be built on the top of various classical CNN backbones. Experimental results demonstrate the superiority of our method. With VGG16 as backbone, we achieve 46.96% Cls-Loc top1 err and 66.12% CorLoc on ILSVRC2014, 38.07% Cls-Loc top1 err and 75.04% CorLoc on CUB200-2011, surpassing the state-of-the-arts by a large margin.
资助项目National Key Research and Development Plan[2018YFC0823003] ; National Key Research and Development Plan[2017YFB1002801] ; Beijing Natural Science Foundation[L182058] ; Natural Science Foundation of China[61876100] ; Natural Science Foundation of China[61972397] ; Shandong Provincial Science and Technology Support Program of Youth Innovation Team in Colleges[2019KJN041]
WOS研究方向Computer Science ; Engineering
语种英语
WOS记录号WOS:000670545900002
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
资助机构National Key Research and Development Plan ; Beijing Natural Science Foundation ; Natural Science Foundation of China ; Shandong Provincial Science and Technology Support Program of Youth Innovation Team in Colleges
源URL[http://ir.ia.ac.cn/handle/173211/45286]  
专题自动化研究所_模式识别国家重点实验室_视频内容安全团队
通讯作者Yuan, Chunfeng
作者单位1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Peking Univ, Sch Software & Microelect, Beijing 100871, Peoples R China
3.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100871, Peoples R China
4.PeopleAI Inc, Beijing 100190, Peoples R China
5.Shandong Technol & Business Univ, Sch Informat & Elect Engn, Yantai 265100, Peoples R China
6.Beijing Inst Tracking & Telecommun Technol BITTT, Beijing 100094, Peoples R China
7.Northwestern Univ, Dept Elect Engn & Comp Sci, Evanston, IL 60208 USA
8.Chinese Acad Sci, CAS Ctr Excellence Brain Sci & Intelligence Techn, Shanghai 200031, Peoples R China
9.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Wang, Bo,Yuan, Chunfeng,Li, Bing,et al. Multi-Scale Low-Discriminative Feature Reactivation for Weakly Supervised Object Localization[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2021,30:6050-6065.
APA Wang, Bo.,Yuan, Chunfeng.,Li, Bing.,Ding, Xinmiao.,Li, Zeya.,...&Hu, Weiming.(2021).Multi-Scale Low-Discriminative Feature Reactivation for Weakly Supervised Object Localization.IEEE TRANSACTIONS ON IMAGE PROCESSING,30,6050-6065.
MLA Wang, Bo,et al."Multi-Scale Low-Discriminative Feature Reactivation for Weakly Supervised Object Localization".IEEE TRANSACTIONS ON IMAGE PROCESSING 30(2021):6050-6065.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。