Fingerprint Recognition Based On Multi-Category SVM

Fingerprint recognition technology is one of the most significant and reliable identification methods as it is unique and invariable.With the development of computer science,fingerprint recongition has extensive outlook servicing people .It is widley used in the field of international trade,criminal,act and a course of justice.So the research of an automatic fingerprint identifiaction system(AFIS)has important technical value and practical value.There are many algorithms in field of fingerprint verication,but they still have some shortages to be researched. This paper studies fingerprint preprocessing ,feature abstraction ,and fingerprint classification.From all of this ,fingerprint classification is the keystone of this paper,and a celerity study anithmetic is used for this classification . The main work of this paper is as follows:1. In the phase of fingerprint preprocessing, conclude and absorb the research result in fingerprint classification anithmetic from home and abroad scholar,implement some important processes ,for example 0-1 process,fine process,feature abstraction process and so on.Finally we gained exactitude iamge data for last celerity identify.2. We test SVM(Support Vector Machine) used two hundred fingerprint images on FVC2004 fingerprint image database .Revelation system capability on different parameter and type.The test result campared between various SVM and Tradition Radial Basic Function Net(RBFN),it revelations excellence generalization capability as classification machine in system anithmetic by us.Especially , C₃-SVM has study celerity rapidness. Experimental result revelation that it has excellence capability in studing speed , classification speed, classification efficiency which we find the fingerprint classification anithmetic system.3. We use Visual C++ 6.0 as soft exploitation tool from Microsofte Company, and OPENCV(Open Source Computation Vision Library )from Intel Company . Finally , we exploitation C3-SVM fingerprint classification system by a aeries of test.Because we design program used face-object method so the system has well expand capability.Fingerprint classification is the main work of this paper.Based on SVM theory it is an exploration in multi-classifiction field.Using these theories ,this paper build a SVM multi-classifiction system .It has been improved better and can work well in practice.