Fingerprint Matching Using User-Defined Markers for Database Filtering

The patent divides the fingerprint matching process into two stages: first, a rapid pre-filtering stage that segments the large database into smaller candidate groups using coarse features; second, a detailed verification stage that applies comprehensive matching only to these reduced candidates. This segmentation maintains final matching accuracy while dramatically reducing overall processing time.

The patent performs preliminary actions by extracting and comparing basic fingerprint features (such as ridge count, pattern type, and minutiae positions) before conducting the full detailed matching process. This preliminary classification reduces the search space, allowing the system to maintain high accuracy through thorough matching of candidates while avoiding time-consuming comparisons with all database entries.

The patent implements feedback mechanisms where the results of preliminary feature comparisons are used to adjust and refine the candidate selection process. The system continuously evaluates matching scores and adjusts the threshold for advancing to detailed verification, ensuring that high-reliability candidates are not missed while maintaining efficient filtering of low-probability matches.

The patent dynamically adjusts matching parameters such as similarity thresholds, candidate selection criteria, and verification depth based on the specific characteristics of the input fingerprint and database conditions. This adaptive parameter adjustment optimizes the balance between reducing candidate numbers for efficiency and maintaining sufficient sensitivity to capture correct matches.