Fingerprint Classification Based on Improved Singular Points Detection and Central Symmetrical Axis

Effective fingerprint classification not only can provide an important index mechanism for large fingerprint database, but also can improve the efficiency and performance of AFIS. At present, because of traditional Poincare method detection more false singular points and weaker anti-noise problem, this paper presents a fingerprint classification method based on continuously directional image and symmetrical axis. Compared with traditional algorithms, this algorithm has the following two aspects improved: firstly, continuously directional image exhibits not only good continuity, well gradualness, and excellent robustness to the noise, but very high precision, which makes singular points location very accurate; secondly, combined singular points quantity and symmetrical axis location relationship divided fingerprint into belonged to classification. Experimental results prove the effectiveness of the algorithm and robustness at Nanjing University fingerprint database and FVC database.