The Untapped Potential of Off-the-Shelf Convolutional Neural Networks

Inkawhich, Matthew; Inkawhich, Nathan; Davis, Eric K.; Li, Hai; Chen, Yiran

doi:10.1109/wacv51458.2022.00296

Cited by 2 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Inkawhich et al present a new adversarial attack based on the modeling of class-wise and layer-wise deep feature distributions (Inkawhich et al 2020b). Inkawhich et al design a flexible attack framework that allows multi-layer perturbations and achieves state-of-the-art targeted transfer-based attack performance (Inkawhich et al 2020a). Yet, this approach requires that the label spaces of the white-box model and the black-box model should overlap.…”

Section: Transfer-based Attackmentioning

confidence: 99%

“…The pre-trained ResNet-50 is provided by the PyTorch library 1 . Note that we do not compare CDTA with relaxed crossdomain attacks, such as FDA+xent (Inkawhich et al 2020a) and No-box Attack (Li, Guo, and Chen 2020). This is because for these methods to work properly, they need to have access to information about the target domain, which is strictly prohibited in this work.…”

Section: Competitorsmentioning

confidence: 99%

See 1 more Smart Citation

CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning

et al. 2023

AAAI

View full text Add to dashboard Cite

Despite the excellent performance, deep neural networks (DNNs) have been shown to be vulnerable to adversarial examples. Besides, these examples are often transferable among different models. In other words, the same adversarial example can fool multiple models with different architectures at the same time. Based on this property, many black-box transfer-based attack techniques have been developed. However, current transfer-based attacks generally focus on the cross-architecture setting, where the attacker has access to the training data of the target model, which is not guaranteed in realistic situations. In this paper, we design a Cross-Domain Transfer-Based Attack (CDTA), which works in the cross-domain scenario. In this setting, attackers have no information about the target model, such as its architecture and training data. Specifically, we propose a contrastive spectral training method to train a feature extractor on a source domain (e.g., ImageNet) and use it to craft adversarial examples on target domains (e.g., Oxford 102 Flower). Our method corrupts the semantic information of the benign image by scrambling the outputs of both the intermediate feature layers and the final layer of the feature extractor. We evaluate CDTA with 16 target deep models on four datasets with widely varying styles. The results confirm that, in terms of the attack success rate, our approach can consistently outperform the state-of-the-art baselines by an average of 11.45% across all target models. Our code is available at https://github.com/LiulietLee/CDTA.

show abstract

Section: Transfer-based Attackmentioning

confidence: 99%

Section: Competitorsmentioning

confidence: 99%

CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning

et al. 2023

AAAI

View full text Add to dashboard Cite

show abstract

“…The classic GAN model has evolved into various forms and architectures, such as DCGAN [ 13 , 14 , 15 , 16 ], Pix2Pix [ 17 , 18 , 19 ], CycleGAN [ 20 , 21 , 22 ], and StyleGAN [ 23 , 24 , 25 , 26 , 27 , 28 ]. There have been several attempts to generate surface defects using these models.…”

Section: Introductionmentioning

confidence: 99%

Image Generation and Recognition for Railway Surface Defect Detection

Xia

Han

Kwon

2023

Sensors

View full text Add to dashboard Cite

Railway defects can result in substantial economic and human losses. Among all defects, surface defects are the most common and prominent type, and various optical-based non-destructive testing (NDT) methods have been employed to detect them. In NDT, reliable and accurate interpretation of test data is vital for effective defect detection. Among the many sources of errors, human errors are the most unpredictable and frequent. Artificial intelligence (AI) has the potential to address this challenge; however, the lack of sufficient railway images with diverse types of defects is the major obstacle to training the AI models through supervised learning. To overcome this obstacle, this research proposes the RailGAN model, which enhances the basic CycleGAN model by introducing a pre-sampling stage for railway tracks. Two pre-sampling techniques are tested for the RailGAN model: image-filtration, and U-Net. By applying both techniques to 20 real-time railway images, it is demonstrated that U-Net produces more consistent results in image segmentation across all images and is less affected by the pixel intensity values of the railway track. Comparison of the RailGAN model with U-Net and the original CycleGAN model on real-time railway images reveals that the original CycleGAN model generates defects in the irrelevant background, while the RailGAN model produces synthetic defect patterns exclusively on the railway surface. The artificial images generated by the RailGAN model closely resemble real cracks on railway tracks and are suitable for training neural-network-based defect identification algorithms. The effectiveness of the RailGAN model can be evaluated by training a defect identification algorithm with the generated dataset and applying it to real defect images. The proposed RailGAN model has the potential to improve the accuracy of NDT for railway defects, which can ultimately lead to increased safety and reduced economic losses. The method is currently performed offline, but further study is planned to achieve real-time defect detection in the future.

show abstract

The Untapped Potential of Off-the-Shelf Convolutional Neural Networks

Cited by 2 publications

References 31 publications

CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning

CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning

Image Generation and Recognition for Railway Surface Defect Detection

Contact Info

Product

Resources

About