Research On Small Area Fingerprint Image Recognition And Matching Algorithm

With the popularity of mobile payments and smart homes,biometrics have gradually captured people's attention.As the first biometric feature applied to safety verification in life,fingerprint recognition is gradually applied in people's daily life.At present,the fingerprint identification algorithm is relatively mature,but with the continuous upgrading of the production process of chips and the like,in various fingerprint products,in order to obtain an aesthetic effect,the collection area of the fingerprint collector becomes smaller and smaller,so the area of the fingerprint image is getting smaller and smaller.Since many fingerprint recognition algorithms rely on the feature points,the detail points of small-area fingerprint images cannot support the algorithm for differential matching.Therefore,the identification and matching of small-area fingerprint images has become a problem to be solved at this stage.Aiming at the above background requirements,this paper proposes a kind of fusion feature,and based on this feature,an identification matching algorithm suitable for small area fingerprint images is constructed.the paper first introduces the basic knowledge of fingerprint recognition,focusing on the characteristics of minutiae and local scale features.The minutiae feature is a common feature unique to fingerprint images.The traditional fingerprint recognition matching algorithm is based on the feature of the detail point.However,as the area of the fingerprint image becomes smaller,the number of minutiae points is reduced,and the minutiae feature cannot be effectively distinguished in matching.In view of the fact that local scale features are local features based on scale space and have been widely used in image stitching and matching,this paper analyzes the fingerprint features based on local scale features by experiment.The experimental results show that the local scale features can be better.to identify the matching fingerprint,but the feature cannot distinguish between the fingerprint image and other images,which is easy to cause false identification.Based on the characteristics of these two features,this paper proposes the fusion feature of the minutiae and the local scale feature.The merged fingerprint feature not only retains the fingerprint specificity of the minutiae point and the specificity of the local scale feature,but also makes full use of the two features.The feature information further enhances the difference of the fingerprint image.In this paper,the fusion process of fingerprint features is introduced in detail,and the parameter of the fusion structure is analyzed in depth.In order to realize the real-time application of large-scale database of fingerprint images and obtain more efficient matching performance,this paper proposes to binarize the fusion structure and introduce the triangular topology to enhance the specificity and robustness of the matching features.According to the characteristics of the fingerprint image,the topology of the triangle is constrained and selected,and finally a high-dimensional feature vector with high specificity and robustness is formed.In this paper,for the two different matching environments of general fingerprint image database and large-scale fingerprint database,the corresponding matching methods are given respectively,and the analysis and suggestions are given.Finally,in order to demonstrate the performance advantages of the proposed fusion feature and the corresponding matching method,combined with the self-built fingerprint image database,the algorithm was tested and analyzed.The experimental results show that the proposed algorithm has better recognition and efficiency than the recently published fingerprint recognition algorithm.