Dual-Decoupling Learning and Metric-Adaptive Thresholding for Semi-supervised Multi-label Learning

被引:0
作者
Xiao, Jia-Hao [1 ]
Xie, Ming-Kun [1 ]
Fan, Heng-Bo [1 ]
Niu, Gang [2 ]
Sugiyama, Masashi [2 ,3 ]
Huang, Sheng-Jun [1 ]
机构
[1] Nanjing Univ Aeronaut & Astronaut, Nanjing, Peoples R China
[2] RIKEN, Ctr Adv Intelligence Project, Tokyo, Japan
[3] Univ Tokyo, Tokyo, Japan
来源
COMPUTER VISION - ECCV 2024, PT LII | 2025年 / 15110卷
基金
国家重点研发计划;
关键词
Multi-label learning; Semi-supervised learning; CLASSIFICATION;
D O I
10.1007/978-3-031-72943-0_25
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Semi-supervised multi-label learning (SSMLL) is a powerful framework for leveraging unlabeled data to reduce the expensive cost of collecting precise multi-label annotations. Unlike semi-supervised learning, one cannot select the most probable label as the pseudo-label in SSMLL due to multiple semantics contained in an instance. To solve this problem, the mainstream method developed an effective thresholding strategy to generate accurate pseudo-labels. Unfortunately, the method neglected the quality of model predictions and its potential impact on pseudo-labeling performance. In this paper, we propose a dual-perspective method to generate high-quality pseudo-labels. To improve the quality of model predictions, we perform dual-decoupling to boost the learning of correlative and discriminative features, while refining the generation and utilization of pseudo-labels. To obtain proper class-wise thresholds, we propose the metric-adaptive thresholding strategy to estimate the thresholds, which maximize the pseudo-label performance for a given metric on labeled data. Experiments on multiple benchmark datasets show the proposed method can achieve the state-of-the-art performance and outperform the comparative methods with a significant margin. The implementation is available at JiahaoXxX/SSMLL-D2L MAT.
引用
收藏
页码:437 / 454
页数:18
相关论文
共 57 条
[1]  
[Anonymous], 2008, SDM
[2]  
[Anonymous], 2009, P ACM INT C IM VID R
[3]   Multi-label Classification with Partial Annotations using Class-aware Selective Loss [J].
Ben-Baruch, Emanuel ;
Ridnik, Tal ;
Friedman, Itamar ;
Ben-Cohen, Avi ;
Zamir, Nadav ;
Noy, Asaf ;
Zelnik-Manor, Lihi .
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022), 2022, :4754-4762
[4]  
Chen Baixu, 2022, ADV NEUR IN
[5]   Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition [J].
Chen, Tianshui ;
Xu, Muxin ;
Hui, Xiaolu ;
Wu, Hefeng ;
Lin, Liang .
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, :522-531
[6]   Multi-Label Image Recognition with Graph Convolutional Networks [J].
Chen, Zhao-Min ;
Wei, Xiu-Shen ;
Wang, Peng ;
Guo, Yanwen .
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, :5172-5181
[7]  
Chinchor N, 1993, 5 MESS UND C MUC 5 P
[8]   Multi-Label Learning from Single Positive Labels [J].
Cole, Elijah ;
Mac Aodha, Oisin ;
Lorieul, Titouan ;
Perona, Pietro ;
Morris, Dan ;
Jojic, Nebojsa .
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, :933-942
[9]   Randaugment: Practical automated data augmentation with a reduced search space [J].
Cubuk, Ekin D. ;
Zoph, Barret ;
Shlens, Jonathon ;
Le, Quoc, V .
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW 2020), 2020, :3008-3017
[10]  
Deng J, 2009, PROC CVPR IEEE, P248, DOI 10.1109/CVPRW.2009.5206848