Research On Image Evaluation And Recognitions Method Of Finger Vein

Biometric recognition refers to the use of distinctive physiological characteristics like fingerprints, face, iris or behavioral characteristics such as gait and signature, for automatically recognizing an individual. In many biometric identification technologies, finger vein recognition technology has become a hot research topic with its unique advantages. Compared with other biometric traits, finger vein patterns have many advantages, such as anti-counterfeiting, user-friendly, liveness and so on. In addition, the finger vein image acquisition device is small. So, it is easy to introduce this technology into a variety of applications.However, in actual situation, there are always a part of poor quality images, because of constantly changing environment, individual differences, and the various performances of devices. These low quality images have a great influence on the performance of finger vein recognition. Therefore, how to effectively evaluate the quality of finger vein image is also a key issue for finger vein recognition. In addition, the research of recognition method has been a hot spot in the field of finger vein recognition. In many recognition methods, superpixel based method of finger vein recognition is a more effective method, but there are still some shortcomings of the existing superpixel based method. Therefore, a new superpixel based finger vein recognition method is presented.We propose a new method based on Support Vector Regression (SVR) for finger vein image quality evaluation. In our method, we use quality scores and quality features to build a SVR model, which will be applied to evaluate quality for testing images. In addition, we use the quality score as soft information to enhance recognition accuracy for finger vein, experimental results show that the recognition performance is improved.In order to make better use of the quality results of finger vein image, we propose an adaptive finger vein recognition framework with image quality analysis. This method can select the appropriate feature extraction method according to the quality difference of the input images, which overcomes the shortcomings of using a fixed feature extraction method. Experimental results show that the proposed framework achieves desirable performance in both speed and recognition rate.There are some shortcomings of the existing superpixel based methods:(1) They can’t distinguished using superpixel (2) There is a lack of more discriminative feature to represent the superpixel. So, a new superpixel based finger vein recognition method is presented. In this method, we choose two types of superpixels, i.e., stable superpixel and discriminative superpixel which will play different roles in matching stage. In detail, the stable and discriminative superpixels are firstly learned from the training images for each enrolled class. When verifying a testing image, we just compare the superpixels at the same location as the two types of superpixels in template. Then we propose a reversible weight based fusion method to combine the two types of superpixels. In addition, in order to improve the recognition performance, we explore the superpixel context feature (SPCF). The experimental results, on two open finger vein databases, show that our proposed method is very effective for finger vein recognition.