Summary of Banktweak: Adversarial Attack Against Multi-object Trackers by Manipulating Feature Banks, By Woojin Shin et al.
BankTweak: Adversarial Attack against Multi-Object Trackers by Manipulating Feature Banks
by Woojin Shin, Donghwa Kang, Daejin Choi, Brent Kang, Jinkyu Lee, Hyeongboo Baek
First submitted to arxiv on: 22 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); 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 Multi-object tracking (MOT) aims to construct moving trajectories for objects, with modern multi-object trackers mainly utilizing the tracking-by-detection methodology. Recent advancements in MOT attacks manipulate object positions to cause persistent identity (ID) switches during association phase, but these position-manipulating attacks have inherent limitations and lack robustness. To improve efficiency and robustness, we present BankTweak, a novel adversarial attack designed for MOT trackers, which focuses on the feature extractor in the association phase and reveals vulnerability in the Hungarian matching method used by feature-based MOT systems. By strategically injecting altered features into the feature banks without modifying object positions, BankTweak induces persistent ID switches even after the attack ends, addressing both efficiency and robustness. We demonstrate the applicability of BankTweak to three multi-object trackers (DeepSORT, StrongSORT, and MOTDT) with one-stage, two-stage, anchor-free, and transformer detectors. Extensive experiments on the MOT17 and MOT20 datasets show that our method substantially surpasses existing attacks, exposing the vulnerability of the tracking-by-detection framework to BankTweak. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary BankTweak is a new way to make it harder for computers to track objects moving in videos. Right now, most object trackers use a method called “tracking by detection” which makes them vulnerable to certain types of attacks that mess with how they work. Some existing attacks try to trick the tracker by changing where the objects are, but these attacks have limitations and aren’t very good at causing trouble. BankTweak is different because it targets the way the tracker figures out what’s happening in each frame, rather than just moving the objects around. This makes it a more efficient and robust attack that can cause problems even after the attack ends. The researchers tested BankTweak on three different types of trackers and showed that it works much better than existing attacks. |
Keywords
» Artificial intelligence » Object tracking » Tracking » Transformer
