AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Unravelling Robustness of Deep Learning Based Face Recognition Against Adversarial Attacks

Gaurav Goswami, Nalini Ratha, Akshay Agarwal, Richa Singh, Mayank Vatsa

AAAI 2018 · Adversarial robustness and security › Black-box and transfer attacks · approach: Robustness benchmarks and defenses · 171 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add grid-based occlusionVGG-Face · PaSCGAR (%) at 1% FAR: 54.3 -> 3.2Hurt
Add most-significant-bit noiseVGG-Face · PaSCGAR (%) at 1% FAR: 54.3 -> 1.3Hurt
Add eye-region occlusionVGG-Face · MEDS-IIGAR (%) at 1% FAR: 78.4 -> 10.9Hurt
Add grid-based occlusionLightCNN · MEDS-IIGAR (%) at 1% FAR: 89.3 -> 80.1Hurt
Add xMSB noiseL-CSSE · MEDS-IIGAR (%) at 1% FAR: 89.1 -> 83.4Hurt
Use hidden-layer response distances with an SVM detectorFace recognition networks · PaSC, proposed image distortionsdetection accuracy (%)Improved
Use hidden-layer response distances with an SVM detectorFace recognition networks · MEDS-II, Gridsdetection accuracy (%): 68.4Mixed
Use only the top 6 hidden-layer features for detection (ablation)VGG-Face detector · MEDS-II, average across distortionsdetection accuracy (%): 81.7%Mixed
Use only the single most discriminative hidden-layer feature (ablation)VGG-Face detector · MEDS-II, average across distortionsdetection accuracy (%): 79.3%Mixed
Apply selective dropout and median filtering to detected distorted imagesLightCNN · PaSCGAR (%) at 1% FAR: 25.9 -> 36.2Improved
Apply selective dropout and median filtering to detected distorted imagesLightCNN · MEDS-IIGAR (%) at 1% FAR: 41.6 -> 61.3Improved
Apply selective dropout and median filtering to detected distorted imagesVGG-Face · PaSCGAR (%) at 1% FAR: 14.6 -> 24.8Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy