Method and device for acquiring direction fields of overlapped fingerprints

Abstract

The invention discloses a method and device for acquiring direction fields of overlapped fingerprints. The method includes the following steps that effective areas of the fingerprints are calibrated, images are partitioned, discrete Fourier transformation is conduced on the partitioned images, low-frequency components are filtered, extreme point pairs of frequency domain amplitude are looked for, and the initial direction fields of the partitioned images are acquired according to the extreme point pairs; matrix modeling is conducted on the initial direction fields of the partitioned images in the overlapped effective areas, relevant matrixes are calculated, initial marked probability matrixes are set, then, random marking, response calculation, marked probability matrix update and convergence judgment are sequentially iteratively conducted, and eventually the separated independent direction fields of the two fingerprints are acquired; the independent effective areas and the separated independent direction fields are combined, and then the separated direction fields of the fingerprints are acquired by selecting one combination mode. The device comprises a calibration module, an initial direction field estimation module, an independent direction field separation module and a combination selection module. The method and device for acquiring the direction fields of the overlapped fingerprints have the advantages of being high in separation quality of the overlapped fingerprints, good in accuracy of ridge detection, and wide in application range.

Description

technical field [0001] The invention relates to the field of fingerprint identification, in particular to a method and a device for obtaining a direction field of overlapping fingerprints. Background technique [0002] Traditional fingerprint recognition systems can recognize fingerprint images that contain only one fingerprint (such as a printed fingerprint) or several non-overlapping fingerprints (such as the entire palm), because traditional fingerprint recognition systems assume that the input fingerprints have been well segmented. However, the emergence of overlapping fingerprints brings challenges to traditional fingerprint identification systems. Reliable ridge detection is very important for successful fingerprint comparison. When the ridge line is clear or the noise of the fingerprint image is small, the ridge line detection method in the prior art can detect the ridge line well, but because the input fingerprint of overlapping fingerprints cannot be segmented, the...

AI Technical Summary

Problems solved by technology

Therefore, how to realize a method of extracting the direction field of overlapping fingerprints on the spot to solve the current problem of being unable to separate overlapping fingerprints, so as to ensure the accuracy of ridge line detection and increase the practicability of the system for automatically separating overlapping fingerprints has become an urgent problem in this field. One of the key technical issues of

Images