CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features

@article{Yun2019CutMixRS,
  title={CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features},
  author={Sangdoo Yun and Dongyoon Han and Seong Joon Oh and Sanghyuk Chun and Junsuk Choe and Youngjoon Yoo},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2019},
  pages={6022-6031}
}
  • Sangdoo Yun, Dongyoon Han, +3 authors Youngjoon Yoo
  • Published 2019
  • Computer Science
  • 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
  • Regional dropout strategies have been proposed to enhance the performance of convolutional neural network classifiers. [...] Key Method We therefore propose the CutMix augmentation strategy: patches are cut and pasted among training images where the ground truth labels are also mixed proportionally to the area of the patches. By making efficient use of training pixels and retaining the regularization effect of regional dropout, CutMix consistently outperforms the state-of-the-art augmentation strategies on…Expand Abstract
    211 Citations
    Attentive Cutmix: An Enhanced Data Augmentation Approach for Deep Learning Based Image Classification
    • 5
    • Highly Influenced
    • PDF
    SmoothMix: a Simple Yet Effective Data Augmentation to Train Robust Classifiers
    • 1
    • Highly Influenced
    • PDF
    SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization
    Regional Patch-Based Feature Interpolation Method for Effective Regularization
    Grad-Cam Guided Progressive Feature CutMix for Classification
    Self-supervised Image Classification based on the Distances of Deep Feature Space
    A visual inductive priors framework for data-efficient image classification.
    GradAug: A New Regularization Method for Deep Neural Networks
    PatchUp: A Regularization Technique for Convolutional Neural Networks
    • 2
    • Highly Influenced
    • PDF
    YOLOv4: Optimal Speed and Accuracy of Object Detection
    • 120
    • Highly Influenced
    • PDF

    References

    SHOWING 1-10 OF 56 REFERENCES
    Improved Regularization of Convolutional Neural Networks with Cutout
    • 711
    • Highly Influential
    • PDF
    LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
    • 570
    • PDF
    DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
    • 5,607
    • PDF
    Attention-Based Dropout Layer for Weakly Supervised Object Localization
    • Junsuk Choe, Hyunjung Shim
    • Computer Science
    • 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
    • 2019
    • 54
    • PDF
    Learning Deep Features for Discriminative Localization
    • 2,712
    • Highly Influential
    • PDF
    RICAP: Random Image Cropping and Patching Data Augmentation for Deep CNNs
    • 22
    • PDF
    Self-produced Guidance for Weakly-supervised Object Localization
    • 67
    • PDF
    Adversarial Complementary Learning for Weakly Supervised Object Localization
    • 164
    • PDF
    Modeling Visual Context is Key to Augmenting Object Detection Datasets
    • 51
    • PDF
    Squeeze-and-Excitation Networks
    • 2,409
    • Highly Influential
    • PDF