Automated Fingerprint Identification Of A Number Of Key Algorithms

Fingerprint recognition has wide application prospect in all fields whichcontain identity authentication. Construction of accurate and reliable, safe andpractical automatic fingerprint identification system (AFIS) has become aresearch hotspot. Although theoretical research and application developmentof AFIS have got a significant progress, accuracy of the algorithm andprocessing speed still need to be improved. In this paper, fingerprint imagepreprocessing algorithms, fingerprint singular points and minutiae extractionalgorithm and fingerprint matching algorithm are analyzed and discussed indetail. Productions are enumerated as follows:1. In the aspect of fingerprint image preprocessing algorithm, firstly,feature vectors composed of Fourier spectrum energy ratio and gray contrastare extracted from fingerprint images sub-blocks and classified by linearsupport vector machine. After classification, morphological operations areperformed to ultimately realize fingerprint image segmentation. Secondly, animproved fingerprint image enhancement algorithm based on Gabor filteringprinciple is presented, which uses adaptive smoothening to estimatefingerprint orientation field and adopts spectrum analysis of projection signalto estimate ridge distance.2. In the aspect of fingerprint feature extraction algorithm, firstly, animproved method of fingerprint singular points extraction is proposed, whichestimates complex square pixel-wise orientation field by multiple scaleslow-pass filtering and takes complex filtering on the orientation field comingwith several heuristic rules for filtering response enhancement. Then, singularpoints' location and orientation are determined according to the responseamplitude and phase. Secondly, an improved minutiae extraction algorithmbased on fingerprint thinned image is presented, which uses open, close andfill operations in fingerprint binary image to avoid producing complicatedfalse minutiae structure and removes false minutiae in the foreground edgeaccording to segmentation mask, and then integrates several kinds ofpost-processing rules to eliminate different types of false minutiae infingerprint thinned image.3. In the aspect of fingerprint feature matching algorithm, a two steps fingerprint matching algorithm, which combines local structure matching andglobal distance matching, is adopted. A kind of minutiae local structure basedon quadrant neighborhood is defined in this algorithm. According to theinstance of fingerprint core point extraction, different searching scheme aboutthe local structure matching is chosen to reduce computation complexity.After the first step matching, central minutiae of the local structure with thegreatest matching score is voted as reference point for features alignment.Finally, elastic matching algorithm is adopted to compute global featuredistance for determination of minutiae's corresponding relation and matchingscore.