Sensitivity of Deep Convolutional Networks to Gabor Noise
Deep Convolutional Networks (DCNs) have been shown to be sensitive to Universal Adversarial Perturbations (UAPs): input-agnostic perturbations that fool a model on large portions of a dataset. These UAPs exhibit interesting visual patterns, but this phenomena is, as yet, poorly understood. Our work shows that visually similar procedural noise patterns also act as UAPs. In particular, we demonstrate that different DCN architectures are sensitive to Gabor noise patterns. This behaviour, its causes, and implications deserve further in-depth study.
Code (1)
Tasks
SensitivitySimilar Papers 제목 키워드 기반
Learnable Gabor modulated complex-valued networks for orientation robustness
Robustness to transformation is desirable in many computer vision tasks, given that input data often exhibits pose variance. While translation invariance and equivariance is a documented phenomenon of CNNs, sensitivity t…
Data AugmentationTranslationGabor is Enough: Interpretable Deep Denoising with a Gabor Synthesis Dictionary Prior
Image processing neural networks, natural and artificial, have a long history with orientation-selectivity, often described mathematically as Gabor filters. Gabor-like filters have been observed in the early layers of CN…
DenoisingDictionary LearningImage DenoisingGaborNet: Gabor filters with learnable parameters in deep convolutional neural networks
The article describes a system for image recognition using deep convolutional neural networks. Modified network architecture is proposed that focuses on improving convergence and reducing training complexity. The filters…
On the Shift Invariance of Max Pooling Feature Maps in Convolutional Neural Networks
This paper focuses on improving the mathematical interpretability of convolutional neural networks (CNNs) in the context of image classification. Specifically, we tackle the instability issue arising in their first layer…
image-classificationImage ClassificationGabor Convolutional Networks
Steerable properties dominate the design of traditional filters, e.g., Gabor filters, and endow features the capability of dealing with spatial transformations. However, such excellent properties have not been well explo…