Fingerprint classification method based on fractal dimension and fingerprint three-level classification method

Abstract

A classification method provided in the invention adopts a step-by-step progressive fingerprint quality assessment algorithm of multiple discrimination factors to reject a low-quality fingerprint so that fingerprint identification accuracy is increased. Three-level classification is performed on the fingerprints with qualified quality. In a first level, according to a fingerprint grain type characteristic, the fingerprints are divided into six types; in a second level, a ridge line number among fingerprint singular points is taken as characteristic so as to carry out classification; and in a third level, a fractal dimension of a fingerprint image quality stabilization area is taken as a characteristic so as to carry out the classification. Through the second level classification and the third level classification, continuous classification and redundant retrieval can be realized so that accuracy and robustness of system classification are effectively increased. An experiment on NIST DB4 shows that the classification of the method provided in the invention is complete; a retrieval speed is fast; an adaptive capacity is high; the high robustness is possessed; and the method is especially suitable for rapid automatic identification of a high capacity fingerprint database.

Description

technical field [0001] The invention relates to the field of fingerprint analysis, more specifically, to a fingerprint classification method based on fractal dimension and a fingerprint three-level classification method. Background technique [0002] Accurate identification of personal identity to ensure information security is a key issue that must be solved in the era of network information. Although the automatic fingerprint identification system, as an effective means of identification, presents a huge development space and application prospect in the fields of public security, finance, e-commerce and personal information security, the current large-capacity fingerprint database automatic fingerprint identification The accuracy and practicability are not satisfactory, and the most critical fingerprint classification problem is a difficult point in the field of pattern recognition. [1 , 2] However, no significant breakthrough has been made, which seriously restricts th...

AI Technical Summary

Problems solved by technology

Most of these methods have two deficiencies: one is that the classification efficiency is reduced due to the uneven distribution probability of fingerprint categories in the natural environment; resulting in non-unique classification results

Images