A ROBUST AUTOMATED FINGERPRINT IDENTIFICATION SYSTEM (AFIS) BASED ON CASCADED QUALITY CHECKS AND MULTIPLE FEATURE SETS AT THREE LEVELS

摘要

The present invention discloses a robust Automated Fingerprint Identification System (AFIS) that makes use of level III features along with Level I (global pattern) & II (minutiae) which significantly improves the system accuracy. This substantial performance is consistent across fingerprint images of varying qualities (good/acceptable/poor). A strict cascaded quality assessment is built into the system for careful inspection and processing of the fingerprint images of varied qualities. The Level III features provide additional discriminatory information, which include but not limited to sweat pores, minutiae constrained ridge contours points, edgeoscopic features and incipient ridges. These extra features are extracted from fingerprint images captured through lOOOppi (or above) resolution fingerprint scanners. This highly customizable and performant system is suitable for both large-scale Civil AFIS and Criminal AFIS.