Font Size: a A A

The Research Of Fingerprint Matching Algorithm And Automatic Fingerprint Identification System

Posted on:2015-06-28Degree:MasterType:Thesis
Country:ChinaCandidate:H L PiFull Text:PDF
GTID:2298330452994438Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
The traditional security system can not meet the need of modern life. The authentication ofelectronic products is in urgent need of a more safer methods. So the technology of fingerprintmatching emerge and develop quickly. Modern science shows that fingerprint recognition hasmany excellent properties, such as stability and uniqueness and has a broad market prospect inthe field of practical applications.The first stage is the pre-processing. The second stage is the feature extraction. Thefeature matching is the third stage. This paper also divide into three parts based production ofpeople:(1)The preprocessing stage includes fingerprint segmentation,fingerprint equilibrium,fingerprint convergence, fingerprint smoothing, fingerprint Intelligence augmentation,fingerprint binarization and denoising,fingerprint thinning and delete the false fingerprintpoints. The experiments show that this algorithm has a good effect lay a solid foundation for thefeature extraction and feature matching。(2)The fingerprint feature extraction section listes a large number of domestic and foreign classical algorithms,the points of center, end and fork,gradient field,,triangle,orientation field are extracted.(3)The image matching algorithm is not quick enough, and the experimental effect is notgood,so I put forward a new matching algorithm here.A fresh feature set, a improved topologyand a new matching process will be used in this fingerprint matching algorithm.To begin with,With the means of K-nearest neighbor method and the ridge line tracing,we will designtopology for fork point and end point respectively.what’s more,we will use Coordinatetransformation, as well as semi-variable bounding box to match every feature point pairs twicerespectively. We hope the process might be finished in advance, but the premise is that thenumber of successful matching feature points must reachthe threshold. The result of fingerprintmatching experiment is very good.The algorithm has a broad market prospect in the field ofpractical applications.
Keywords/Search Tags:Fingerprint matching, K-nearest neighbor method, the ridge line tracing, semi-variable bounding box
PDF Full Text Request
Related items