Rethinking the Localization in Weakly Supervised Object Localization

被引:2
|
作者
Xu, Rui [1 ]
Luo, Yong [2 ,3 ]
Hu, Han [4 ]
Du, Bo [2 ,3 ]
Shen, Jialie [5 ]
Wen, Yonggang [6 ]
机构
[1] Wuhan Univ, Sch Comp Sci, Wuhan, Peoples R China
[2] Wuhan Univ, Wuhan, Peoples R China
[3] Hubei Luojia Lab, Wuhan, Peoples R China
[4] Beijing Inst Technol, Beijing, Peoples R China
[5] City Univ London, London, England
[6] Nanyang Technol Univ, Singapore, Singapore
基金
新加坡国家研究基金会; 中国国家自然科学基金;
关键词
weakly supervised; object localization; binary-class detector; weighted entropy; noisy label;
D O I
10.1145/3581783.3611959
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Weakly supervised object localization (WSOL) is one of the most popular and challenging tasks in computer vision. This task is to localize the objects in the images given only the image-level supervision. Recently, dividing WSOL into two parts (class-agnostic object localization and object classification) has become the state-of-the-art pipeline for this task. However, existing solutions under this pipeline usually suffer from the following drawbacks: 1) they are not flexible since they can only localize one object for each image due to the adopted single-class regression (SCR) for localization; 2) the generated pseudo bounding boxes may be noisy, but the negative impact of such noise is not well addressed. To remedy these drawbacks, we first propose to replace SCR with a binary-class detector (BCD) for localizing multiple objects, where the detector is trained by discriminating the foreground and background. Then we design a weighted entropy (WE) loss using the unlabeled data to reduce the negative impact of noisy bounding boxes. Extensive experiments on the popular CUB-200-2011 and ImageNet-1K datasets demonstrate the effectiveness of our method.
引用
收藏
页码:5484 / 5494
页数:11
相关论文
共 50 条
  • [41] Adaptively Denoising Proposal Collection for Weakly Supervised Object Localization
    Wenju Xu
    Yuanwei Wu
    Wenchi Ma
    Guanghui Wang
    Neural Processing Letters, 2020, 51 : 993 - 1006
  • [42] Combinational Class Activation Maps for Weakly Supervised Object Localization
    Yang, Seunghan
    Kim, Yoonhyung
    Kim, Youngeun
    Kim, Changick
    2020 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV), 2020, : 2930 - 2938
  • [43] Weakly-supervised object localization in unlabeled image collection
    Qu, Yanyun
    Liu, Han
    Yang, Xiaoqing
    Fang, Suwen
    Wang, Hanzi
    MULTIMEDIA SYSTEMS, 2013, 19 (01) : 51 - 63
  • [44] Task-Aware Weakly Supervised Object Localization With Transformer
    Meng, Meng
    Zhang, Tianzhu
    Zhang, Zhe
    Zhang, Yongdong
    Wu, Feng
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (07) : 9109 - 9121
  • [45] Weakly Supervised Learning for Object Localization Based on an Attention Mechanism
    Park, Nojin
    Ko, Hanseok
    APPLIED SCIENCES-BASEL, 2021, 11 (22):
  • [46] Discovering an inference recipe for weakly-supervised object localization
    Lee, Sanghuk
    Mun, Cheolhyun
    Uh, Youngjung
    Choe, Junsuk
    Byun, Hyeran
    PATTERN RECOGNITION, 2024, 156
  • [47] Weakly Supervised Object Localization Using Things and Stuff Transfer
    Shi, Miaojing
    Caesar, Holger
    Ferrari, Vittorio
    2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2017, : 3401 - 3410
  • [48] IMPROVING CLASS ACTIVATION MAP FOR WEAKLY SUPERVISED OBJECT LOCALIZATION
    Zhang, Zhenfei
    Chang, Ming-Ching
    But, Tien D.
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 2624 - 2628
  • [49] Diverse Complementary Part Mining for Weakly Supervised Object Localization
    Meng, Meng
    Zhang, Tianzhu
    Yang, Wenfei
    Zhao, Jian
    Zhang, Yongdong
    Wu, Feng
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2022, 31 : 1774 - 1788
  • [50] Sparse weakly supervised models for object localization in road environment
    Zadrija, Valentina
    Krapac, Josip
    Segvic, Sinisa
    Verbeek, Jakob
    COMPUTER VISION AND IMAGE UNDERSTANDING, 2018, 176 : 9 - 21