Corpus ID: 208158412

AdvSPADE: Realistic Unrestricted Attacks for Semantic Segmentation

@article{Shen2019AdvSPADERU,
  title={AdvSPADE: Realistic Unrestricted Attacks for Semantic Segmentation},
  author={Guangyu Shen and Chengzhi Mao and Junfeng Yang and Baishakhi Ray},
  journal={arXiv: Computer Vision and Pattern Recognition},
  year={2019}
}
  • Guangyu Shen, Chengzhi Mao, +1 author Baishakhi Ray
  • Published 2019
  • Computer Science, Engineering
  • arXiv: Computer Vision and Pattern Recognition
  • Due to the inherent robustness of segmentation models, traditional norm-bounded attack methods show limited effect on such type of models. In this paper, we focus on generating unrestricted adversarial examples for semantic segmentation models. We demonstrate a simple and effective method to generate unrestricted adversarial examples using conditional generative adversarial networks (CGAN) without any hand-crafted metric. The na\"ive implementation of CGAN, however, yields inferior image… CONTINUE READING
    1 Citations
    Multitask Learning Strengthens Adversarial Robustness
    • 6
    • PDF

    References

    SHOWING 1-10 OF 50 REFERENCES
    On the Robustness of Semantic Segmentation Models to Adversarial Attacks
    • 109
    • Highly Influential
    • PDF
    Adversarial Examples for Semantic Segmentation and Object Detection
    • 382
    • PDF
    A Unified Gradient Regularization Family for Adversarial Examples
    • 97
    • PDF
    Detecting Adversarial Samples from Artifacts
    • 416
    • PDF
    High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs
    • 1,292
    • PDF
    Generating Adversarial Examples with Adversarial Networks
    • 250
    • Highly Influential
    • PDF
    Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
    • 674
    • PDF
    Houdini: Fooling Deep Structured Prediction Models
    • 152
    • PDF
    Image-to-Image Translation with Conditional Adversarial Networks
    • 7,142
    • PDF