Design of an enhancement algorithm for low quality fingerprint evidence images
During the process of fingerprint image acquisition,there are inherent differences such as rotation,translation,and incomplete image acquisition.The number of fine nodes extracted each time is not fixed,and this feature form has a high computa-tional complexity for later matching.The current image enhancement algorithms based on template matching aim to address the high computational complexity of low-quality fingerprint evidence image matching,and complete the original image filtering process with a fixed threshold without considering the issue of single pixel thinning direction field feature loss.Propose a low quality finger-print evidence image enhancement method based on Kalman filtering.Based on the Gaussian noise distribution function of the im-age,VisuShrink threshold denoising is selected to construct a Kalman filter and complete fingerprint image preprocessing.Convert the original image into a single pixel refined image,and extract the singular points and feature points of the filtered fingerprint im-age.Divide fingerprint images,determine the direction field of the fingerprint based on singular points and direction vectors,per-form image enhancement filtering processing,and complete the design of a low-quality fingerprint evidence image enhancement method based on Kalman filtering.Construct an experimental section to verify the application effect of this method.The experimen-tal method shows that after using this method,low-quality fingerprint images become clearer,and the average error rate is only 6.21%,with an average evaluation score of 0.85.This method effectively improves the image enhancement effect.
Kalman filterfingerprint enhancementimage enhancementdirection fieldfeature point extractionimage preprocessing
NETL
NSTL
万方数据
