| Traditional security system is mainly based on token or password. With the development of society, this system is becoming more fragile. To cope with this challenge, we look to biometrics and hope to enhance identity verification by using our body's physical character and behavior.A fingerprint is the pattern of ridges and valleys on the surface of a fingertip. Modern anatomy and statistics have proved that each individual has unique fingerprints and fingerprint doesn't change all the life time, i. e. fingerprint is steady and exclusive. As a primary biometric, fingerprint has been widely used in forensic, police and other security system and been studied by using pattern recognition technique firstly. Fingerprint identification has been the pronoun of biometrics in many situations.Automatic fingerprint identification system (AFIS) has integrated image processing, pattern recognition and database technique. Tracing back to 1960's when it's first introduced (is it right?), AFIS has been successful and still the hotspot in pattern recognition domain today after 40 years' developmentBased on plenty of papers and technical reports on fingerprint matching and classification, this dissertation provides some works on the theories on AFIS, including two main aspects which are fingerprint matching and classification, respectively. Furthermore, a set of measurement dealing with fingerprint matching and classification captured by solid-state sensors is proposed. Through experiments, this measure is proved rational and practical. The main points are as follows:Chapter 1 introduces some typic biometrics such as fingerprint, face, iris, handprint, signature, voice print and so on. We have compared these biometrics on seven aspects e.g. stability, uniqueness, performance and collectability and presented two working modes and how to evaluate biometrics system. Although it has been reproached by some lawyers in the past ?, its' reliability is doubtless as a tested science by practice. This chapter also includes the structure of AFIS and some popular sensors used in the system. Later on functions and composing of classification subsystem, matching subsystem and compression subsystem are listed. In the last part of this chapter, some questions embarrassing??? the development of AFIS are presented on fingerprint capturing, image enhancement, representation, classification and compression.Chapter 2 mainly discusses preprocessing of fingerprint image captured by solid-state sensors. Above all, a new, fast quality assessing algorithm is presented by analyzing direction of fingerprint. By this means, we can estimate whether the impression is too small, or the finger is partial, or the finger is wet or dry. Secondly, we synthesize a few measurements to divide the image into unrecoverable regions and recoverable regions. By using FFT, the recoverable regions are enhanced and separate and average filters are used to eliminate false forks and ruptures. The enhanced image is delivered to binarization module. A new binarization and filtering algorithm based on direction is used to gain favorable impact. Thinning is the key to extract minutiae. Based on classical algorithm, this dissertation proposed a method of the nearest neighbour to satisfy the requirement of preserve, connect and speediness.Chapter 3 introduces a way how to construct integrated template of registration and how to match fingerprints by using integrated template. So called "integrated template", means that we take into account not only the location, direction and type of minutiae, but also its' local texture feature and confidence. So we present how to extract minutiae in preprocessed fingerprint image and how to prune minutiae in order to reserve the most credible ones based on topology and distributing rules first of all. Considering that point pattern matching is indispensable to making integrated template and matching fingerprints, we also proposed a fast minutiae matching algorithm based on clustering. By this method we first try to find the p...
|
PDF Full Text Request