Adaptive Calibration for Fair and Performant Facial Recognition
2026-06-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors present Adaptive Calibration, a new way to make facial recognition systems more accurate and fair. Their method adjusts how similarity scores between face features are translated into probabilities based on local context, fixing a problem where the same score might mean different things in different cases. This approach improves both overall accuracy and fairness without needing information about people's demographics. It works better than previous methods and avoids harming some groups to help others.
Facial recognitionCalibrationCosine similarityEmbeddingsProbability mappingFairnessDemographic metadataMachine learning benchmarksAccuracy metricsPretrained models
Authors
Ryan Brown, Chris Russell
Abstract
We introduce Adaptive Calibration (AC), a novel calibration strategy for facial recognition that maps cosine similarity between normalized embeddings to well-calibrated probabilities. By incorporating local context into calibration, Adaptive Calibration corrects for a fundamental mismatch in cosine similarity, whereby the same distance can correspond to different match probabilities in different embedding regions. Our approach improves both overall performance and results in a fairer calibration without requiring demographic metadata. Our approach consistently dominates existing methods both on accuracy and fairness metrics across a variety of pretrained models and standard benchmarks. AC provides a practical solution for equitable facial recognition, without requiring demographic group annotations, and while improving overall performance. Unlike existing approaches, our method provides continuous, region-specific calibration that avoids "leveling down" where fairness comes at the cost of degraded performance for some groups.