Research On Latent Fingerprint Preprocessing Technique Based On Partial Differential Equation Inpainting

Latent fingerprint left at crime scenes has played a prominent role in the field of criminal investigation as a special kind of important evidence, so it has an extremely important research value. With the rapid development of science and technology, living-fingerprint recognition technology has made great progress. But due to some unique characteristics of the latent fingerprints such as ambiguity, fragmentary nature, complex background, ridge line discontinuities, ridge overlapping, it is still in a bottleneck state so far to find an effective latent fingerprints enhancement technology.Image restoration is one of the most classical problems in image processing. Some objects which are not expected in the image, such as unconscious scratch, signature and so on, can be effectively reduced by image restoration techniques. The performance of fingerprint recognition system mainly depends on the quality of fingerprint image, while the images collected have some degrees of fracture and scratches, this requires us prior to repair before the processing of fingerprint enhancement, it not only ensures the continuity of the enhanced ridge, but also reduce the number of redundant eigen points after the refinement and improve the matching rate and recognition rate.Therefore, a new method of restoration is proposed for partially fouling, breaking and scratch latent fingerprint.Aiming at the generally acknowledged problem of latent fingerprint is difficult to accurate authenticate, this paper proposes a critical step of fingerprint restoration in preprocessing procedure. It is necessary for partially fouling, breaking and scratch latent fingerprint. According to research Partial Differential Equation, such fingerprint composed by the local parallel ridge and valley should be difficult to be repaired with general restoration algorithm. Based on partial differential equations algorithm, this paper presents a fingerprint restoration algorithm composing gradient and orientation field. According to gradient and orientation field of the known pixel points, different weights are used in different orientation field in the restoration process. Then Gabor filter algorithm is adopted to enhance fingerprint image, fingerprint processing realized by using binaryzation thinning and feature point extraction algorithm. Experimental results demonstrate that the proposed restoration algorithm can effectively reduce the false feature points. Finally, the improved algorithm is compared with the former, shows that the feasibility, effectiveness and superiority of this algorithm.