Research On Fast Finding Method Of Weak Random Statistical Feature Point Set Fingerprint Database

With the widespread application of automatic fingerprint identification technology,the capacity of the corresponding fingerprint database is constantly increasing.How to quickly and accurately search for large-scale fingerprint data is a hot issue for fingerprint identification technology.This paper selects the characteristics of the fingerprint details and proposes a fast fingerprint database retrieval method to solve the problem of fast searching for large-capacity fingerprint databases.Based on the weak randomness between each detail point of the fingerprint,this paper extracts reliable retrieval features,and proposes a fast fingerprint database retrieval method based on the core detail point support system.First,the fingerprint image is converted into a set of detail point vectors,each detail point is taken as the core detail point,and all the detail points except the current core detail point are used as supporting detail points,the core detail point and its supporting details.The points form the core detail point support system.Secondly,a suitable indexer is designed according to the constructed retrieval features.This paper uses the feature vectors formed by supporting details to construct a two-dimensional hash table.The indexer constructed using hashing technology can help us quickly in the retrieval stage.Estimate the number of supporting detail point pairs that were successfully matched.Third,the global best registration method is applied to find the true matching detail point pairs from the estimated detail point pairs,thereby improving the accuracy of the retrieval score.Finally,the feature vector of a retrieved fingerprint is given during retrieval,and the fingerprint template with the retrieved feature vector in the database can be obtained directly through hash table addressing.By counting the number of occurrences,the fingerprint can be found and retrieved.Similar library fingerprint templates.The experimental results show that the fingerprint retrieval system based on the position and orientation information of the fingerprint detail points proposed in this paper has a better retrieval rate and accuracy than other types of fingerprint retrieval systems based on the features of the detail points.