Study On Quality Assessment Algorithm Of Finger Vein Image

A large number of scholars and researchers in many areas focus on the quality assessment algorithm,especially in the field related to biometric identification technology,the quality assessment algorithm is well applied.In biometric identification technology,the finger vein recognition technology is a hot topic,many scholars have made deep research on this.They found that if the finger vein image quality is too low,the recognition accuracy of the finger vein recognition system will be reduced,which would influence the identification process.The low quality of finger vein image is caused by many reasons,such as uneven illumination,changing of the temperature and humidity,the position of the finger placed during the process of capturing images,etc.Therefore,the paper put forward a kind of algorithm,with which the high quality and low quality of finger vein image would be effectively distinguished to solve the above problems,based on the analysis of the finger vein image.Firstly,on the basis of collecting images of finger vein,this paper has analyzed the various factors that affect images and the characteristics of the images from different aspects,and extracted six different characteristic parameters,such as the spatial domain gradient,the contrast,the two-dimensional entropy of image,the position deviation degree,the signal-to-noise ratio and the effective area.Then,two methods have used to set up the finger vein image quality evaluation model through the processing of the characteristic parameters.The first one method is the finger vein image quality assessment model based on the improved weighted fusion algorithm.The model is established based on the directly weighted fusion of the first five kinds of characteristic parameters,and takes the influence of different characteristics on finger vein images into consideration.The experimental results show that the high and low quality finger vein images can be classified effectively by this model,which can solve the problem of the paper.The second method is the finger vein image quality assessment model based on the support vector machine(SVM)classification algorithm.This method has finally chosen the three characteristic parameters of the gradient in the spatial domain,contrast and the effective area as the input characteristics of the model through the study of the experiment of different characteristic parameters.A 3D reference frame has been built using the three characteristics.The radial basis function(RBF)has been selected as the kernel function of support vector machine(SVM)model,which made the linearly inseparable problem into a separable one.Then theSVM quality evaluation model has be set up.The experimental results have shown that the SVM quality evaluation model established could effectively classify the finger vein images and also could improve the recognition performance of recognition system to some extent.