Research On Dry Fingerprint Image Enhancement And Feature Extraction

With the development of the information industry,the demand for biometric recognition technology of identity authentication is increasingly urgent.Fingerprint recognition technology is one of the most widely used and mature biometric recognition technologies.At present,the fingerprint recognition technology that processes good fingerprint images is relatively mature.However,in the public security system,due to the dry fingerprint image formed by the dry skin of the fingers,the ridge line structure is incomplete and the contrast is low.The traditional fingerprint recognition technology is not effective.This paper takes the dry fingerprint image as the research object and studies its key algorithm in the recognition system.It mainly includes: Improved acquisition,segmentation,enhancement and feature extraction of dry fingerprint images.The main work can be summarized as follows.(1)Research on the acquisition of dry fingerprint images.According to the relationship between the humidity of dry fingerprint image and light illumination,this paper presents an automatic optimization method for fingerprint image acquisition.It improves the quality of dry fingerprint images,lays the foundation for the enhancement and feature extraction of dry fingerprint images.(2)Studies fingerprint segmentation algorithm.This paper analysises the advantages and disadvantages of traditional segmentation algorithms.And presents an improved segmentation algorithm which combines Canny operator with morphology.Firstly,It uses the improved Canny operator to perform edge of dry fingerprint images.Secondly,a series of morphological filters are performed on the effective edges of the foreground using appropriate structural elements.Finally,the edge of the fingerprint is smoothly filtered.Experimental results show,this algorithm has higher accuracy and effective information at the edge of the fingerprint is not losed.(3)Studies directional filtering and Gabor filtering enhancement algorithm.The Gabor filter enhancement algorithm is improved both on the ridge direction and frequency acquisition method.Firstly,use an adaptive method to set the size of the image block to obtain relatively accurate direction parameters.Secondly,calculate the average frequency of fingerprint local area.Finally,a variable-direction filter is used in conjunction with the direction and frequency of the ridges to enhance dry fingerprint images.Experimental results show,the algorithm has a better enhancement effect and consumes less time.(4)Studies fingerprint Image Feature Extraction and eliminate false minutiaes.Use the adaptive threshold method to binarize and fingerprint image and use improved OPTA thinning algorithm to thin the image.Extract the endpoints and bifurcation points of the thinning image.This paper uses different algorithms to eliminate false minutiaes.The deletion rate of false features is high,which improves the accuracy of recognition.