FNMR

FNMR stands for False Non-Match Rate, a biometric performance metric used to evaluate verification systems. It measures how often a genuine user is incorrectly rejected when attempting to authenticate. In practice, FNMR indicates the frequency with which legitimate attempts fail to produce a match.

FNMR is computed for a given decision threshold as the number of genuine comparisons that do not yield a match divided by the total number of genuine attempts. In many publications FNMR is used interchangeably with the false rejection rate (FRR) at a particular threshold. The Genuine Accept Rate (GAR) is the complement of FNMR at that same threshold (GAR = 1 − FNMR).

FNMR is commonly analyzed alongside other biometric metrics such as the false acceptance rate (FAR). There is often a trade-off between FNMR and FAR: lowering the threshold to reduce FNMR can increase FAR, and vice versa. Performance is frequently summarized using DET curves or ROC-style analyses to show how FNMR (or FRR) varies with FAR across operating points.

Factors affecting FNMR include the quality of the biometric sample (signal clarity, noise, and alignment), sensor resolution and condition, environmental factors (lighting, distance), user-specific factors (age, wearing conditions), and enrollment quality. FNMR is relevant across biometric modalities such as fingerprint, iris, facial recognition, and voice-based systems, and it is a key consideration in designing and evaluating secure and user-friendly verification systems.