NLM filtering finger vein denoising method based on skin crack segmentation

Abstract

The invention discloses an NLM filtering finger vein denoising method based on skin crack segmentation. The method comprises: firstly, using a multi-scale Frangi filtering algorithm for carrying out filtering response analysis on a finger vein image with skin crack characteristics caused by finger ecdysis, and selecting the spatial scale size and a segmentation threshold value; performing NLM filtering denoising on the detected area interfered by the skin cracks according to the skin crack binary image, and not performing denoising processing on the non-interfered area. The finger vein denoising algorithm based on switch type non-local mean filtering performs denoising specifically aimed at the exuvial skin crack area, pseudo vein interference is reduced, loss of information of a normal vein area is avoided, redundant information of other areas on the image can be fully utilized when the method is applied to the finger vein image, a certain restoration effect can be achieved, and the denoising effect is better. Therefore, the method is a finger vein denoising algorithm which is good in finger ecdysis type finger vein image denoising effect.

Description

technical field [0001] The invention belongs to the technical field of biological feature identification and information security, and in particular relates to the field of finger vein image preprocessing. Background technique [0002] Finger vein technology, as the latest biometric technology, has been applied in various fields such as banking, access control punching system, and social security by virtue of its high anti-counterfeiting performance, high accuracy, and rapid recognition. A complete finger vein recognition system, including image acquisition, image preprocessing, feature extraction and matching recognition. The image preprocessing part includes region of interest extraction, size grayscale normalization, image denoising, image enhancement, etc. Image preprocessing determines the quality of the extracted features and has a huge impact on subsequent matching recognition, so it is very important. However, most of the finger vein collection devices currently pop...

AI Technical Summary

Problems solved by technology

[0004] For the skin crack area in the vein image of finger molt, it is easy to produce pseudo veins to affect the recognition performance, and the existing finger vein denoising algorithm is easy to cause the overall blurring of the image and the loss of vein information. The present invention provides a segmentation method based on skin cracks. The non-local mean (Non-Local Means, NLM) filtering finger vein denoising method

Images