Inferior-quality fingerprint grayscale image enhancement method on basis of three gaussian filtering

Abstract

The invention discloses an inferior-quality fingerprint grayscale image enhancement method on the basis of three gaussian filtering. The method is used for studying parameter values in a three-gaussian model of a non-canonical receptive field in a human visual system according to an image characteristic that fingerprint ridges and furrows are alternately distributed and through introducing the three-gaussian model of the non-canonical receptive field in the human visual system into fingerprint grayscale image processing. The method comprises the following steps of: according to the size of a fingerprint grayscale image, obtaining fingerprint characteristic parameters; carrying out three-gaussian one-sided filtering on the fingerprint grayscale image to obtain neighborhood subjective feeling brightness of the fingerprint grayscale image; and according to a difference between the obtained neighborhood subjective feeling brightness and an actual light intensity, carrying out linear local contrast adjustment on the fingerprint grayscale image. The method disclosed by the invention can be used for effectively improving the overall and local contrast and brightness and highlighting the detail characteristics of dark areas, thus being especially applicable to the processing on inferior-quality fingerprint grayscale images.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a low-quality fingerprint grayscale image processing method combining human visual characteristics and fingerprint features. Background technique [0002] There are two types of fingerprint grayscale image enhancement methods: pixel-based image enhancement and texture-based image enhancement. The former is to process the fingerprint in the airspace, and directly acts on the pixel value of the fingerprint. It is generally used in the primary processing of the fingerprint, and mainly realizes the adjustment of the image gray level, so that the subsequent extraction of the direction map and frequency map of the fingerprint gray image is more accurate. Favorable, thus laying a good foundation for the subsequent further enhancement of fingerprints by using direction or frequency information. The latter is carried out in the transform domain, using the texture information of ...

AI Technical Summary

Problems solved by technology

However, in the field of criminal investigation, low-quality fingerprint grayscale images are usually collected at the crime scene. The main reasons are: discontinuous ridges, that is, there are interruptions or cracks in the ridges; ridges caused by sweat pores, stains, etc. Line adhesion Parallel ridges cannot be separated well; there are creases, abrasions or trauma on the fingerprint; deformation due to uneven pressure and angular rotation during collection

[0004] The above factors can easily lead to the loss of real features, false features in the extracted features, and errors in the position features (position and direction) of minutiae points, etc.

Therefore, for low-quality fingerprint grayscale images, it is difficult to correctly obtain the orientation map or frequency map of the fingerprint, and the automatic fingerprint recognition system is basically not competent, and more manual participation is required

Images