LocalScore: Local Density-Aware Similarity Scoring for Biometrics

2026-02-01

AI Summary

This work addresses the challenge in open-set biometrics where unenrolled probes are difficult to detect due to conventional methods compressing intra-class variations into a single global representation, resulting in ambiguous decision boundaries and poor robustness. To overcome this limitation, the authors propose LocalScore, a plug-and-play algorithm that introduces, for the first time in biometric scoring, an explicit k-nearest neighbor–based local density estimation to construct a similarity scoring mechanism sensitive to the local structure of the feature space. LocalScore incurs minimal computational overhead and is compatible with arbitrary model architectures and loss functions. Experimental results demonstrate significant improvements in open-set performance across multiple modalities: the open-set retrieval FNIR@FPIR drops from 53% to 40%, and verification TAR@FAR increases from 51% to 74%.

Technology Category

• Search and Optimization: Local Search
• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Biometrics, Face, Gesture & Pose

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Open-set biometrics faces challenges with probe subjects who may not be enrolled in the gallery, as traditional biometric systems struggle to detect these non-mated probes. Despite the growing prevalence of multi-sample galleries in real-world deployments, most existing methods collapse intra-subject variability into a single global representation, leading to suboptimal decision boundaries and poor open-set robustness. To address this issue, we propose LocalScore, a simple yet effective scoring algorithm that explicitly incorporates the local density of the gallery feature distribution using the k-th nearest neighbors. LocalScore is architecture-agnostic, loss-independent, and incurs negligible computational overhead, making it a plug-and-play solution for existing biometric systems. Extensive experiments across multiple modalities demonstrate that LocalScore consistently achieves substantial gains in open-set retrieval (FNIR@FPIR reduced from 53% to 40%) and verification (TAR@FAR improved from 51% to 74%). We further provide theoretical analysis and empirical validation explaining when and why the method achieves the most significant gains based on dataset characteristics.

Problem

• open-set biometrics
• non-mated probes
• multi-sample galleries
• intra-subject variability
• local density

Innovation

• LocalScore
• local density
• open-set biometrics
• k-nearest neighbors
• multi-sample gallery