Towards Robust Fine-grained Recognition by Maximal Separation of Discriminative Features

06/10/2020
by   Krishna Kanth Nakka, et al.
0

Adversarial attacks have been widely studied for general classification tasks, but remain unexplored in the context of fine-grained recognition, where the inter-class similarities facilitate the attacker's task. In this paper, we identify the proximity of the latent representations of different classes in fine-grained recognition networks as a key factor to the success of adversarial attacks. We therefore introduce an attention-based regularization mechanism that maximally separates the discriminative latent features of different classes while minimizing the contribution of the non-discriminative regions to the final class prediction. As evidenced by our experiments, this allows us to significantly improve robustness to adversarial attacks, to the point of matching or even surpassing that of adversarial training, but without requiring access to adversarial samples.

READ FULL TEXT

page 20

page 21

page 22

page 24

page 25

page 26

page 27

page 28

research
09/06/2019

Coarse2Fine: A Two-stage Training Method for Fine-grained Visual Classification

Small inter-class and large intra-class variations are the main challeng...
research
12/01/2020

Robustness Out of the Box: Compositional Representations Naturally Defend Against Black-Box Patch Attacks

Patch-based adversarial attacks introduce a perceptible but localized ch...
research
10/26/2022

Improving Adversarial Robustness with Self-Paced Hard-Class Pair Reweighting

Deep Neural Networks are vulnerable to adversarial attacks. Among many d...
research
05/30/2022

Exposing Fine-grained Adversarial Vulnerability of Face Anti-spoofing Models

Adversarial attacks seriously threaten the high accuracy of face anti-sp...
research
10/14/2015

Fine-Grained Product Class Recognition for Assisted Shopping

Assistive solutions for a better shopping experience can improve the qua...
research
10/15/2019

Understanding Misclassifications by Attributes

In this paper, we aim to understand and explain the decisions of deep ne...
research
03/15/2023

Improving Adversarial Robustness with Hypersphere Embedding and Angular-based Regularizations

Adversarial training (AT) methods have been found to be effective agains...

Please sign up or login with your details

Forgot password? Click here to reset