GB/Z 153-2025 Information technology—Biometrics—Characterization and measurement of difficulty for fingerprint databases for technology evaluation

Scope

This document provides guidance on estimating the challenge and pressure level of fingerprint datasets for fingerprint recognition based on factors such as relative sample quality, relative rotation, deformation, and fingerprint overlap. Additionally, it establishes methods for constructing datasets of different difficulty levels and defines the relative difficulty levels of fingerprint datasets for technical evaluation of fingerprint recognition algorithms. The difficulty level depends on the differences between the reference samples and probe samples in the aforementioned factors. This document includes the following content:
—— Describes the difficulty level caused by differences between different samples collected from the same finger;
—— Summarizes the influencing factors and establishes a statistical method for representing the difficulty level of the entire fingerprint dataset;
—— Compares the difficulty levels of different fingerprint datasets;
—— Defines the testing and reporting process for the difficulty level of fingerprint datasets used for technical evaluation;
—— Analyses the data characteristics of paired samples based on comparison scores;
—— Describes the method of selecting archived data for establishing evaluation datasets.
This document provides guidance on comparing the relative difficulty levels of fingerprint datasets.
This document does not include:
—— Defining the quality of a single fingerprint image;
—— Defining methods or clear indicators for evaluating or predicting the performance of fingerprint recognition algorithms.

Development Information

Adopt standards

Related Standards