跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.21923 · v1

Wrong Input, Wrong Threshold: Re-evaluating Adversarial Facial Privacy

首次提交 2026-11-17 05:37
版本更新 2026-11-17 05:37
雷达观测 2026-11-18 19:43
arXiv ↗PDF ↗

摘要

Published face-privacy methods report protection above 90% against black-box recognition models. That number comes from an evaluation protocol with two independent defects, reported in a measure that cannot size either one. The input is wrong: methods published since 2022 feed the network the whole photograph, or a detector's box, where deployed systems detect, align to a landmark template and crop. Re-scoring the same files through a realistic front end costs a standard transfer attack 56% of its dodging protection and a budget-allocating attack 7%, with no image regenerated. The measure is wrong: that loss is a difference of success rates, so it saturates, moving by 8.2x over a budget sweep while the effect beneath it moves by 1.34x. We therefore report the paired cosine shift, in which both thresholds cancel algebraically, so the input defect holds without committing to a calibration. Measured that way it is 2.2x and 10.7x larger on three unseen recognizers, and at matched budget it reverses the L-infinity transfer ranking: the one surveyed method that aligns at evaluation gains 0.18 where MI-FGSM loses 0.05. The cause is a resampling mismatch, not perturbation lost outside the crop; optimising through a differentiable alignment operator largely removes it, adding +24 to +58 points on unseen recognizers and a commercial system, with a residual on one of them. The thresholds are wrong: the four constants inherited since 2021 hold the false-accept rate near its nominal 1% on aligned crops but admit 2.3-5.4% on whole images, with an opposite-signed bias across dodging and impersonation, the two tasks reported side by side. Differentiable alignment and gallery scoring are not new: the first generation of this literature used both, and we show what dropping them cost.

作者备注:10 pages main text and 15 pages supplementary material. Submitted to IEEE Transactions on Information Forensics and Security. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-17 05:37 · 最新观测版本

Wrong Input, Wrong Threshold: Re-evaluating Adversarial Facial Privacy

Published face-privacy methods report protection above 90% against black-box recognition models. That number comes from an evaluation protocol with two independent defects, reported in a measure that cannot size either one. The input is wrong: methods published since 2022 feed the network the whole photograph, or a detector's box, where deployed systems detect, align to a landmark template and crop. Re-scoring the same files through a realistic front end costs a standard transfer attack 56% of its dodging protection and a budget-allocating attack 7%, with no image regenerated. The measure is wrong: that loss is a difference of success rates, so it saturates, moving by 8.2x over a budget sweep while the effect beneath it moves by 1.34x. We therefore report the paired cosine shift, in which both thresholds cancel algebraically, so the input defect holds without committing to a calibration. Measured that way it is 2.2x and 10.7x larger on three unseen recognizers, and at matched budget it reverses the L-infinity transfer ranking: the one surveyed method that aligns at evaluation gains 0.18 where MI-FGSM loses 0.05. The cause is a resampling mismatch, not perturbation lost outside the crop; optimising through a differentiable alignment operator largely removes it, adding +24 to +58 points on unseen recognizers and a commercial system, with a residual on one of them. The thresholds are wrong: the four constants inherited since 2021 hold the false-accept rate near its nominal 1% on aligned crops but admit 2.3-5.4% on whole images, with an opposite-signed bias across dodging and impersonation, the two tasks reported side by side. Differentiable alignment and gallery scoring are not new: the first generation of this literature used both, and we show what dropping them cost.

查看此版本解读 →查看此版本原文 ↗