Adversarial Metric Attack and Defense for Person Re-Identification
- 16 October 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 43 (6) , 2119-2126
Abstract
Person re-identification (re-ID) has attracted much attention recently due to its great importance in video surveillance. In general, distance metrics used to identify two person images are expected to be robust under various appearance changes. However, our work observes the extreme vulnerability of existing distance metrics to adversarial examples, generated by simply adding human-imperceptible perturbations to person images. Hence, the security danger is dramatically increased when deploying commercial re-ID systems in video surveillance. Although adversarial examples have been extensively applied for classification analysis, it is rarely studied in metric analysis like person re-identification. The most likely reason is the natural gap between the training and testing of re-ID networks, that is, the predictions of a re-ID network cannot be directly used during testing without an effective metric. In this work, we bridge the gap by proposing Adversarial Metric Attack, a parallel methodology to adversarial classification attacks. Comprehensive experiments clearly reveal the adversarial effects in re-ID systems. Meanwhile, we also present an early attempt of training a metric-preserving network, thereby defending the metric against adversarial attacks. At last, by benchmarking various adversarial settings, we expect that our work can facilitate the development of adversarial attack and defense in metric-based applications.Keywords
All Related Versions
Funding Information
- Engineering and Physical Sciences Research Council (Seebibyte EP/M013774/1)
- EPSRC/MURI (EP/N019474/1)
This publication has 44 references indexed in Scilit:
- Scalable Person Re-identification: A Benchmark2015
- Person Re-Identification Ranking Optimisation by Discriminant Context Information Analysis2015
- Person re-identification by Local Maximal Occurrence representation and metric learning2015
- Similarity learning on an explicit polynomial kernel feature map for person re-identification2015
- High-Speed Tracking with Kernelized Correlation FiltersIEEE Transactions on Pattern Analysis and Machine Intelligence, 2014
- DeepReID: Deep Filter Pairing Neural Network for Person Re-identification2014
- DeepFace: Closing the Gap to Human-Level Performance in Face Verification2014
- Reidentification by Relative Distance ComparisonIEEE Transactions on Pattern Analysis and Machine Intelligence, 2012
- Large scale metric learning from equivalence constraints2012
- Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features2008