A Self-supervised GAN for Unsupervised Few-shot Object Recognition

被引:1
|
作者
Nguyen, Khoi [1 ]
Todorovic, Sinisa [1 ]
机构
[1] Oregon State Univ, Corvallis, OR 97330 USA
关键词
D O I
10.1109/ICPR48806.2021.9412539
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper addresses unsupervised few-shot object recognition, where all training images are unlabeled, and test images are divided into queries and a few labeled support images per object class of interest. The training and test images do not share object classes. We extend the vanilla GAN with two loss functions, both aimed at self-supervised learning. The first is a reconstruction loss that enforces the discriminator to reconstruct the probabilistically sampled latent code which has been used for generating the "fake" image. The second is a triplet loss that enforces the discriminator to output image encodings that are closer for more similar images. Evaluation, comparisons, and detailed ablation studies are done in the context of few-shot classification. Our approach significantly outperforms the state of the art on the Mini-Imagenet and Tiered-Imagenet datasets.
引用
收藏
页码:3225 / 3231
页数:7
相关论文
共 50 条
  • [31] Self-supervised few-shot medical image segmentation with spatial transformations
    Titoriya, Ankit Kumar
    Singh, Maheshwari Prasad
    Singh, Amit Kumar
    [J]. Neural Computing and Applications, 2024, 36 (30) : 18675 - 18691
  • [32] Self-Supervised Traffic Classification: Flow Embedding and Few-Shot Solutions
    Horowicz, Eyal
    Shapira, Tal
    Shavitt, Yuval
    [J]. IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT, 2024, 21 (03): : 3054 - 3067
  • [33] A Novel Self-supervised Few-shot Network Intrusion Detection Method
    Zhang, Jing
    Shi, Zhixin
    Wu, Hao
    Xing, Mengyan
    [J]. WIRELESS ALGORITHMS, SYSTEMS, AND APPLICATIONS (WASA 2022), PT I, 2022, 13471 : 513 - 525
  • [34] Self-Supervised and Few-Shot Contrastive Learning Frameworks for Text Clustering
    Shi, Haoxiang
    Sakai, Tetsuya
    [J]. IEEE ACCESS, 2023, 11 : 84134 - 84143
  • [35] Decoupled Self-Supervised Subspace Classifier for Few-Shot Class-Incremental SAR Target Recognition
    Zhao, Yan
    Zhao, Lingjun
    Zhang, Siqian
    Liu, Li
    Ji, Kefeng
    Kuang, Gangyao
    [J]. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2024, 17 : 15845 - 15861
  • [36] Few-shot defect recognition of metal surfaces via attention-embedding and self-supervised learning
    Liu, Zhenyu
    Song, Yiguo
    Tang, Ruining
    Duan, Guifang
    Tan, Jianrong
    [J]. JOURNAL OF INTELLIGENT MANUFACTURING, 2023, 34 (08) : 3507 - 3521
  • [37] Few-shot defect recognition of metal surfaces via attention-embedding and self-supervised learning
    Zhenyu Liu
    Yiguo Song
    Ruining Tang
    Guifang Duan
    Jianrong Tan
    [J]. Journal of Intelligent Manufacturing, 2023, 34 : 3507 - 3521
  • [38] SFT: Few-Shot Learning via Self-Supervised Feature Fusion With Transformer
    Lim, Jit Yan
    Lim, Kian Ming
    Lee, Chin Poo
    Tan, Yong Xuan
    [J]. IEEE ACCESS, 2024, 12 : 86690 - 86703
  • [39] Meta Self-Supervised Learning for Distribution Shifted Few-Shot Scene Classification
    Gong, Tengfei
    Zheng, Xiangtao
    Lu, Xiaoqiang
    [J]. IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, 2022, 19
  • [40] Self-supervised pairwise-sample resistance model for few-shot classification
    Weigang Li
    Lu Xie
    Ping Gan
    Yuntao Zhao
    [J]. Applied Intelligence, 2023, 53 : 20661 - 20674