论文标题
使用低通激活功能和DCT增强功能的强大图像分类
Robust Image Classification Using A Low-Pass Activation Function and DCT Augmentation
论文作者
论文摘要
卷积神经网络（CNN）的性能差异最近受到了审查。在这项工作中，我们分析了频域中的常见损坏，即高频损坏（HFC，例如噪声）和低频损坏（LFC，例如Blur）。尽管对HFC的简单解决方案是低通滤波，但是Relu-广泛使用的激活函数（AF）没有任何过滤机制。在这项工作中，我们将低通滤网灌输到AF（LP-RELU）中，以提高针对HFC的鲁棒性。为了处理LFC，我们将LP-Relu与基于离散余弦变换的增强相辅相成。 LP-Relu再加上DCT扩展，使一个深层的网络可以解决整个腐败范围。我们使用CIFAR-10-C和Tiny Imagenet-C进行评估，并且与最先进的（SOTA）相比，准确性分别提高了5％和7.3％。我们进一步评估了我们的方法在CIFAR-10-P和Tiny Imagenet-P中的各种扰动上的稳定性，在这些实验中也实现了新的SOTA。为了进一步加强我们对CNN缺乏鲁棒性的理解，在这项工作中提出了决策空间可视化过程。
Convolutional Neural Network's (CNN's) performance disparity on clean and corrupted datasets has recently come under scrutiny. In this work, we analyse common corruptions in the frequency domain, i.e., High Frequency corruptions (HFc, e.g., noise) and Low Frequency corruptions (LFc, e.g., blur). Although a simple solution to HFc is low-pass filtering, ReLU -- a widely used Activation Function (AF), does not have any filtering mechanism. In this work, we instill low-pass filtering into the AF (LP-ReLU) to improve robustness against HFc. To deal with LFc, we complement LP-ReLU with Discrete Cosine Transform based augmentation. LP-ReLU, coupled with DCT augmentation, enables a deep network to tackle the entire spectrum of corruption. We use CIFAR-10-C and Tiny ImageNet-C for evaluation and demonstrate improvements of 5% and 7.3% in accuracy respectively, compared to the State-Of-The-Art (SOTA). We further evaluate our method's stability on a variety of perturbations in CIFAR-10-P and Tiny ImageNet-P, achieving new SOTA in these experiments as well. To further strengthen our understanding regarding CNN's lack of robustness, a decision space visualisation process is proposed and presented in this work.
