Research And Implementation Of A Fingerprint Identification System Based On Classification Features

With the advent of the information age,biometric recognition technology has been flourishing,and has achieved some research and application results in the field of personal identity authentication.Fingerprint identification technology,with its characteristics of easy collection,stable identification and low cost,has become the highest proportion of identity authentication technology in biometric recognition technology industry.Due to the huge population base,the research of fingerprint recognition technology has gradually shifted to automatic fingerprint recognition based on large-capacity fingerprint database.In order to find a fingerprint template that matches the fingerprint to be recognized,the fingerprint database needs to establish an effective index,otherwise it will be scanned in the database with a large time cost,which does not meet the requirements of people for an efficient system.Therefore,this paper takes the classification method in fingerprint recognition system as the main research direction,studies and realizes the three-level fingerprint classification algorithm and the fingerprint recognition system based on classification features,and reduces the entire fingerprint database to a homologous fingerprint set which issimilar to the fingerprint to be identified.(1)In the stage of fingerprint image preprocessing,a set of fingerprint quality evaluation method is studied and implemented,which can eliminate the low-quality fingerprint and prompt the user to collect again.The evaluation method adopts the characteristics of effective area ratio,effective gray contrast,effective area centroid offsetand so on.(2)In terms of fingerprint classification,on the one hand,the three-level fingerprint classification method is studied.Firstly,the fingerprint pattern is determined according to the number and position of the singular points which are extracted,and the fingerprint is roughly classified.Secondly,according to the number of ridge lines between singular points to achieve secondary classification.Finally,the classification is based on the average period of ridges.The three-feature data are used to build an indexed large-capacity fingerprint database.On the other hand,in view of the inaccurate estimation of the number of ridges between singular points and the average periodof ridges,the database index method is studied and implemented.When the matching of homologous fingerprint sets fails,the search radius is increased.Through a large number of experiments,the upper limit of the search radius is obtained,which reduces the search time and rejection rate of the system.Based on the existing algorithm combined with the above two aspects of research content,a set of fingerprint automatic identification system for postgraduate enrollment examination is implemented.The experimental data show that the quality evaluation method and classification algorithm can reduce the rejection rate of the system and improve the efficiency of the system.