Research On Fingerprint Identification Technology Based On PCNN And Fuzzy Vault Algorithms

Fingerprint identification technology is a technology which uses the fingerprint characteristics to verify the identification, and is the most mature and most promising biometric technology. It contains four parts: fingerprint image access, image processing, feature extraction and preservation, and matching. Based on studying traditional fingerprint identification technology, an improved binary fingerprint image thinning algorithm using template-based PCNNs and an improved approach to generated Fuzzy Vault based on the pseudo-random sequence are proposed. Both algorithms test on FVC2004 database. Moreover, a visualization image interface is designed in Visual C++ 6.0, which can be applied to fingerprint image processing.An improved binary fingerprint image thinning algorithm using template-based PCNNs is a method that is based on parallel capability of PCNN after giving square- and-triangle-template. The proposed algorithm iteratively deletes border pixels, in which the combined sequential thinning using square templates and parallel thinning using triangle templates are employed to accelerate execution. The results show that the method can effectively delete border pixels without breaking the original connectivity, overcome not thorough thinning, and reduce breaks and spikes, and can apply to the automatic fingerprint recognition system to improve accept rate. The dissertation introduces this algorithm as well as a mathematical model of template-based PCNNs.A Novel approach to generated fuzzy vault based on the pseudo-random sequence is a method that the vectors of the fingerprint templates can be transformed by the defined sequence, which makes them more random and improves the safety of key. A bounding box of variable size minutiae matcher is applied during decoding to account for nonlinear distortion and this leads to find out reliable minutiae to improve the genuine accept rate. The experimental results demonstrate that we can achieve revocability or modification of vector, hide the key and defend against many attacks. Besides, it can enhance the complexity of key cracked.The fingerprint identification system consisting of the above two improved algorithms can be accurately extracted fingerprint feature points to improve the matching precision and can be more secure to manage fingerprint template database.