Table 5 Comparison of Adversarial Accuracy (Adv Acc) for various adversarial methods. A higher Adversarial Accuracy indicates a less effective adversarial manipulation, whereas a lower Adversarial Accuracy signifies greater system vulnerability.
From: Evaluating gait system vulnerabilities through PPO and GAN-generated adversarial attacks
Method | Adversarial Accuracy (%) |
|---|---|
Adversarial attack using GANs (2019)50 | 80 |
Gait spoofing using masterization (2023)54 | 78 |
Adversarial attacks using accelerometer-based data (2022)52 | 78.9 |
Temporal sparse adversarial attack using GAN (2023)33 | 67.78 |
One cycle attack on gait authentication (2021)51 | 57 |
Reinforcement learning-based Q-learning attack (2023)53 | 47.96 |
Infrared Adversarial Patch Generation Based on Reinforcement Learning (2024)55 | 37.70 |
Adversarial attack using GAN & PPO (Our Method) | 31.40 |
