Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FACTUAL: A Novel Framework for Contrastive Learning Based Robust SAR Image Classification

Xu Wang, Tian Ye, Rajgopal Kannan, Viktor Prasanna

TL;DR

SAR ATR models are vulnerable to adversarial manipulation, including physically feasible attacks. FACTUAL combines supervised adversarial contrastive pre-training with perturbations from PGD and OTSA to learn robust, class-aware representations, followed by fine-tuning a linear classifier. The approach yields high accuracy on clean data and strong robustness against unseen attacks, evidenced by TA ≈ 99.7% and RA ≈ 89.6% on MSTAR with a small TA−RA gap, outperforming prior methods. This work advances robust SAR classification by integrating labeled contrastive learning with realistic adversarial perturbations, supporting practical deployment in all-weather, all-time SAR scenarios.

Abstract

Deep Learning (DL) Models for Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), while delivering improved performance, have been shown to be quite vulnerable to adversarial attacks. Existing works improve robustness by training models on adversarial samples. However, by focusing mostly on attacks that manipulate images randomly, they neglect the real-world feasibility of such attacks. In this paper, we propose FACTUAL, a novel Contrastive Learning framework for Adversarial Training and robust SAR classification. FACTUAL consists of two components: (1) Differing from existing works, a novel perturbation scheme that incorporates realistic physical adversarial attacks (such as OTSA) to build a supervised adversarial pre-training network. This network utilizes class labels for clustering clean and perturbed images together into a more informative feature space. (2) A linear classifier cascaded after the encoder to use the computed representations to predict the target labels. By pre-training and fine-tuning our model on both clean and adversarial samples, we show that our model achieves high prediction accuracy on both cases. Our model achieves 99.7% accuracy on clean samples, and 89.6% on perturbed samples, both outperforming previous state-of-the-art methods.

FACTUAL: A Novel Framework for Contrastive Learning Based Robust SAR Image Classification

TL;DR

SAR ATR models are vulnerable to adversarial manipulation, including physically feasible attacks. FACTUAL combines supervised adversarial contrastive pre-training with perturbations from PGD and OTSA to learn robust, class-aware representations, followed by fine-tuning a linear classifier. The approach yields high accuracy on clean data and strong robustness against unseen attacks, evidenced by TA ≈ 99.7% and RA ≈ 89.6% on MSTAR with a small TA−RA gap, outperforming prior methods. This work advances robust SAR classification by integrating labeled contrastive learning with realistic adversarial perturbations, supporting practical deployment in all-weather, all-time SAR scenarios.

Abstract

Deep Learning (DL) Models for Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), while delivering improved performance, have been shown to be quite vulnerable to adversarial attacks. Existing works improve robustness by training models on adversarial samples. However, by focusing mostly on attacks that manipulate images randomly, they neglect the real-world feasibility of such attacks. In this paper, we propose FACTUAL, a novel Contrastive Learning framework for Adversarial Training and robust SAR classification. FACTUAL consists of two components: (1) Differing from existing works, a novel perturbation scheme that incorporates realistic physical adversarial attacks (such as OTSA) to build a supervised adversarial pre-training network. This network utilizes class labels for clustering clean and perturbed images together into a more informative feature space. (2) A linear classifier cascaded after the encoder to use the computed representations to predict the target labels. By pre-training and fine-tuning our model on both clean and adversarial samples, we show that our model achieves high prediction accuracy on both cases. Our model achieves 99.7% accuracy on clean samples, and 89.6% on perturbed samples, both outperforming previous state-of-the-art methods.
Paper Structure (17 sections, 7 equations, 2 figures, 2 tables)

This paper contains 17 sections, 7 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Adversarial Attacks
  4. Contrastive Learning
  5. Method
  6. Overall Framework
  7. Data Augmentation
  8. Supervised Adversarial Contrastive Pre-training
  9. Supervised Adversarial Fine-tuning
  10. Experiment
  11. Dataset
  12. Attacks
  13. Model
  14. Metrics
  15. Experiment Design
  16. ...and 2 more sections

Figures (2)

  • Figure 1: Framework architecture of FACTUAL. The region with blue dashed line as border refers to the Data Augmentation step. The region with green dashed line as border refers to the Supervised Adversarial Contrastive Pre-training step. The region with Red dashed line as border refers to the Supervised Adversarial Fine-tuning step.
  • Figure 2: Clean sample and perturbed samples of SAR image