Summary of Rethinking Impersonation and Dodging Attacks on Face Recognition Systems, by Fengfan Zhou et al.
Rethinking Impersonation and Dodging Attacks on Face Recognition Systems
by Fengfan Zhou, Qianyu Zhou, Bangjie Yin, Hui Zheng, Xuequan Lu, Lizhuang Ma, Hefei Ling
First submitted to arxiv on: 17 Jan 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed Adversarial Pruning (Adv-Pruning) method is designed to enhance both impersonation and dodging attacks on Face Recognition (FR) systems simultaneously. This novel approach consists of Priming, Pruning, and Restoration stages. By leveraging Adversarial Priority Quantification and Biased Gradient Adaptation, Adv-Pruning prioritizes features that maintain original perturbation effectiveness while boosting dodging capabilities. Compared to state-of-the-art methods, comprehensive experiments demonstrate the superiority of Adv-Pruning in achieving successful dodging attacks on FR systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Adversarial Pruning is a new way to trick Face Recognition (FR) systems. The goal is to make fake faces that can deceive these systems without being detected. Current methods are good at making fake faces that impersonate specific people, but they’re not very good at hiding from the system. This method makes both types of fake faces and fine-tunes them to get better results. It does this by prioritizing certain features in the fake face images and adapting them to trick the FR system. |
Keywords
* Artificial intelligence * Boosting * Face recognition * Pruning
