Research On Low Quality Fingerprint Matching Algorithm Based On Small Area Phase Correlation

The method based on detail point features is the common method in fingerprint matching,which mostly needs to find the best correspondence of detail points for matching.However,in the process of detail point extraction,the number of extracted detail points is not enough to satisfy the matching requirements due to the low-quality of fingerprint images,causing this method being not suitable for the matching of low-quality fingerprint images.In order to solve the problem of detail point matching of low-quality fingerprint images,this paper proposes a small-area phase-related low-quality fingerprint matching algorithm.The method uses the improved hill climbing search method to calculate the best phase correlation between the small-area fingerprint template and the fingerprint image search to be matched,so as to realize fingerprint matching.First,for sake of overcoming the difficulty of the proposed algorithm applying to multiple scenarios,the URU4000 B collector,FVC2004 data set B library scale fingerprint image and SOCOFing Dataset Obl degree for easy,medium,and hard fingerprint images as an experimental dataset.Secondly,the fingerprint image in the data set is evaluated in terms of quality and then the evaluated image will undergo the preprocessing,which includes: fingerprint image segmentation,correction,enhancement and small-area template selection.Among them,the purpose of fingerprint image segmentation is to remove the excess background and better retain the foreground.The fingerprint image correction can effectively unify the fingerprint image specifications,which is conducive to the subsequent matching.The process is that first Fourier transform the segmented fingerprint image,and then binary transform.Next,transform the value by the Hough amplitude.Moreover,take the Hough transform line and finally correct the fingerprint image according to the direction Angle obtained by the Hough transform;Fingerprint image enhancement effectively solves the definition of the ridge line in the effective area of the fingerprint image,which is conducive to the subsequent low-quality small area fingerprint template selection and fingerprint matching.The selection of low-quality fingerprint image small area fingerprint template effectively selects the small area fingerprint template in the matching process to facilitate the subsequent improvement of phase correlation matching of hill climbing search.Finally,the accuracy of the proposed algorithm was verified through numerous simulation experiments,effectively solving the problem of difficult matching of low-quality fingerprint images.The experimental results show that the proposed algorithm is particularly effective in matching low-quality fingerprint images with poor matching performance.From the perspective of the peak value of the phase correlation function,the proposed algorithm improves the peaks of pure phase correlation and traditional uphill search small-area phase correlation by 0.472 and0.509,respectively.From the perspective of matching degree,compared with the algorithms based on fine node features,pure phase correlation and traditional hill climbing search small-area phase correlation,the proposed algorithm improves the matching degree by 19.4,11.8 and 16.6 percentage points,respectively.