Research And Design Of Automatic Fingerprint Identification System

Fingerprint identification is one of the most widely used biometric technology in practice,with the development of computer and information technology,the research and application of automatic fingerprint identification technology is currently one of the hot research topic of biometric recognition,has been widely used in all kinds of personal identification and verification system.Automatic fingerprint recognition technology can be divided into two main processes,fingerprint image preprocessing and fin gerprint matching.In this paper,through the analysis of fingerprint recognition algorithm in each step of the fingerprint image quality dependence,summed up the necessity of adding fingerprint enhancement model;through different types of analysis of fi ngerprint image noise,the fingerprint enhancement model is proposed to enhance the specific requirements.To this end,this paper proposes a fingerprint enhancement model,in view of the different regions of the fingerprint image characteristics of differ ent frequency domain,select the appropriate filter to enhance,and ultimately to achieve the overall enhancement effect.After background segmentation of fingerprint images,the model first calculates the consistency of the direction field.By direction field consistency information and image segmentation mask,we can locate the singular point area which needs special processing.For the non singular point region,the average frequency of the ridge line is used as the frequency parameter of the Gabor filter,because of the small variation of the ridge line frequency.Because the tiny difference in the frequency of the fingerprint image thinning operation will no longer be reflected,so the calculation of the exact frequency of each pixel is not significant.For the singular point region,this paper analyzes the characteristics of the spectrum map,and select the appropriate bandwidth filter to filter.By adjusting the bandwidth of the filter to get better filtering effect,and no longer dependent on the direction of the region is not reliable.Such processing can reduce the error of the directional information brought about by the generation of the pseudo ridge structure.The experimental results show that the model can effectively remove most of the fingerprint image noise and improve the accuracy of minutiae extraction steps,which has practical significance for the fingerprint automatic recognition system to improve the robustness.In the fingerprint image enhancement based on the model,combining the fingerprint image phase matching algorithm and the classical fingerprint matching algorithm based on image features,using Microsoft C#.NET to realize the automatic fingerprint identification system in detail based on feature matching,the experiment shows that the system has good usability.