ComDefend: An Efficient Image Compression Model to Defend Adversarial Examples
By Xiaojun Jia, Xingxing Wei, Xiaochun Cao, Hassan Foroosh. Computer Vision and Pattern Recognition, 2018
331 citations recorded by the source corpus.
Deep neural networks (DNNs) have been demonstrated to be vulnerable to adversarial examples. Specifically, adding imperceptible perturbations to clean images can fool the well trained deep neural networks. In this paper, we propose an end-to-end image compression model to defend adversarial examples: ComDefend. The proposed model consists of a compression convolutional neural network (ComCNN) and a reconstruction convolutional neural network (ResCNN). The ComCNN is used to maintain the structure information of the original image and purify adversarial perturbations. And the ResCNN is used to reconstruct the original image with high quality. In other words, ComDefend can transform the adversarial image to its clean version, which is then fed to the trained classifier. Our method is a pre-processing module, and does not modify the classifier’s structure during the whole process. Therefore it can be combined with other model-specific defense models to jointly improve the classifier’s robustness. A series of experiments conducted on MNIST, CIFAR10 and ImageNet show that the proposed method outperforms the state-of-the-art defense methods, and is consistently effective to protect classifiers against adversarial attacks.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition
View Paper: https://www.semanticscholar.org/paper/33e3ea69244ac597d61b9e31098b02fbfccd086d
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Ajmal Saeed Mian
- Mubarak Shah
- Navid Kardan
- Naveed Akhtar
- Liwei Wang
- Kun He
- John E. Hopcroft
- Jiadong Lin
In Computer Vision and Pattern Recognition
- Ajmal Saeed Mian
- Mubarak Shah
- Luc Van Gool
- Dacheng Tao
- Xiaogang Wang
- Naveed Akhtar
- Andrew Zisserman
- Bernt Schiele
Citations in the corpus
- Nesterov Accelerated Gradient and Scale Invariance for Adversarial Attacks (2019, methodology): “…guided denoiser (HGD) (Liao et al., 2018), random resizing and padding (R&P) (Xie et al., 2018), NIPS-r3 1 , feature distillation (FD) (Liu et al., 2019), purifying perturbations via image compression model (Comde-fend) (Jia et al., 2019) and randomized smoothing (RS) (Cohen et al., 2019).”
- A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, unknown): “Adversarial examples [96], [97], [98], [99], [100], [101], [102], [103], [104], [105] also involve the adversarial idea.”
- Advances in adversarial attacks and defenses in computer vision: A survey (2021, methodology): “Instead of directly using standard image compression, learnable compression methods that use neural models are also proposed in the literature for adversarial defense [397], [398].”