False Non-Match Rate (FNMR)
Definition
The proportion of mate comparisons (same individual) that the system incorrectly scores as non-matching. Also called the false negative rate. A lower FNMR means fewer failures to recognise the same person.
- Also called
- False negative rate (biometric matching)
- Formula
- FNMR = false non-matches / total mate comparisons
- Consequence
- Legitimate user or true match rejected by the system
- Tradeoff partner
- False Match Rate (FMR), set by the match threshold
Common questions
What raises FNMR in a real deployment?+
Poor image quality, pose or lighting change, aging between the reference and probe images, occlusion, and a threshold set conservatively to keep FMR low all push FNMR up. It is a system-and-conditions number, not a fixed property of an algorithm.
Why report FNMR alongside FMR rather than either alone?+
A single error rate can be made to look good by shifting the threshold, which moves the other rate in the opposite direction. Reporting both, ideally at a stated operating threshold, is what lets one system be compared fairly against another.
Related terms
- ArcFace / FaceNet
- Two influential deep CNN facial recognition architectures. FaceNet (Google, 2015) used a triplet-loss training approach to learn facial embeddings. ArcFace (Deng et...
- Embedding
- A compact numerical vector representation of a face produced by a neural network. The distance between two embeddings in the vector space...
- False Match Rate (FMR)
- The proportion of non-mate comparisons (different individuals) that the system incorrectly scores as a match. Also called the false positive rate or...
- Gallery vs Probe
- In facial recognition, the gallery is the reference database of enrolled face images; the probe is the query image being compared against...
- NIST FRVT
- The National Institute of Standards and Technology Face Recognition Vendor Testing programme, the primary independent benchmark for facial recognition algorithms, publishing ongoing...