Embedding verification method controls false acceptance using identity ranking
Identities, Not Pairs: Distribution-Free False-Accept Control for Embedding-Based Verification
Machine Learning
Summary
When systems check if two images or data points are from the same person, they need to avoid wrongly accepting impostors. The authors found that setting thresholds based on pairs can be misleading because some identities behave like hubs, skewing results. They designed a new method called COVE that ranks the test pair among all possible pairs to control false acceptance strictly. Tests showed COVE keeps false acceptance low and true acceptance rates high, improving verification accuracy.
What this means in practice
- •For biometric system developers: Improve verification accuracy and false accept control in biometric face or fingerprint authentication by applying COVE's identity-based thresholding method.
- •For security software engineers: Enhance access control systems by reliably calibrating acceptance thresholds to reduce false acceptance without sacrificing true accepts using the proposed ranking technique.
Authors
Salim Khazem
Abstract
Embedding-based verification often calibrates a false-accept threshold from all impostor pairs among $m$ identities, although these pairs share identities. We characterize a worst-case limit: when hub identities score highly against a fraction $ρ$ of the population, a calibration-only threshold with marginal false-accept rate (FAR) at most $α$ under every identity-exchangeable law reaches the hub score on hub-free calibration sets with probability at most $(m+2)α/(2ρ)$. Uniformly accepting near that score at levels well below $ρ/m$ therefore requires trial information. We propose COVE, which ranks the trial pair among all $N=\binom{m+2}{2}$ pairs formed by augmenting the calibration units with the probe and reference. Under identity exchangeability, its marginal FAR is at most $\lfloorαN\rfloor/N$, with first possible acceptance at $1/N=O(1/m^2)$. The rule also retains this guarantee with a fixed symmetric cohort-normalized score using an independent cohort. On 6 benchmarks and 8 frozen encoders, realised FAR was consistent with the exact bound. Where both valid calibration-only baselines accepted nothing, COVE retained a median $100\%$ of the sampled oracle's true-accept rate over reliable comparisons. In contrast, an empirical quantile over dependent pairs reached 2.87$\,α$ FAR, and a binomial upper confidence bound over pairs missed its nominal 95\% conditional-FAR target in up to 55\% of calibration draws. Symmetric CSLS raised COVE's true-accept rate by up to 4.3-fold in the tested settings. The code is available in https://github.com/salimkhazem/Cove.git