Anti-aliasing deep image classifiers using novel depth adaptive blurring and activation function
- Authors: Hossain, Md Tahmid , Teng, Shyh , Lu, Guojun , Rahman, Mohammad Arifur , Sohel, Ferdous
- Date: 2023
- Type: Text , Journal article
- Relation: Neurocomputing Vol. 536, no. (2023), p. 164-174
- Full Text: false
- Reviewed:
- Description: Deep convolutional networks are vulnerable to image translation or shift, partly due to common down-sampling layers, e.g., max-pooling and strided convolution. These operations violate the Nyquist sampling rate and cause aliasing. The textbook solution is low-pass filtering (blurring) before down-sampling, which can benefit deep networks as well. Even so, non-linearity units, such as ReLU, often re-introduce the problem, suggesting that blurring alone may not suffice. In this work, first, we analyse deep features with Fourier transform and show that Depth Adaptive Blurring is more effective, as opposed to monotonic blurring. To this end, we propose a novel Depth Adaptive Blur-pool (DAB-pool) module to replace existing down-sampling methods. Second, we introduce a novel activation function – with a built-in low pass filter, as an additional measure, to keep the problem from reappearing. From experiments, we observe generalisation on other forms of transformations and corruptions as well, e.g., rotation, scale, and noise. We evaluate our method under three challenging settings: (1) a variety of image translations; (2) adversarial attacks – both
Robust image classification using a low-pass activation function and DCT augmentation
- Authors: Hossain, Md Tahmid , Teng, Shyh , Sohel, Ferdous , Lu, Guojun
- Date: 2021
- Type: Text , Journal article
- Relation: IEEE Access Vol. 9, no. (2021), p. 86460-86474
- Full Text:
- Reviewed:
- Description: 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. © 2013 IEEE.