Research On Distributed Dorsal Hand Vein Image Quality Assessment

With the rapid development of information technology, Internet technology become a "double-edged sword", which cause information leakage increasingly becoming the focus of attention. The traditional identity authentication modes including password and key can not satisfy the needs of current society. So biometric technology, which overcomes the inconvenience of traditional modes, with the stability, uniqueness, portability and other advantages, has become a hot research and occupied more and more market share. And fingerprint, face and iris have become common means of identification in our lives.Dorsal hand vein recognition, which has following features:anti-counterfeiting, high difficult to dirty and high acceptance, has paid greater attention from scholars around the world in recent years. Because of the close relation between image quality and the performance of biometric identification system, poor sample quality is a main reason for matching errors in the identifying process of biometric systems. Therefore it is necessary to introduce the quality evaluation process. This paper introduces dorsal hand vein image acquisition device and database. Then stepwise quality evaluation and multi-parameter fusion algorithm were used for quality evaluation respectively. At last, the problems of heterogeneity hand vein images which caused by distributed systems were studied.The innovative work of this paper can be summarized as follows:(1) According to the reference standards for biometric image quality assessment and the characters of dorsal hand vein image, For dorsal hand vein, some performance metrics, brightness, contrast, the size of image effective region, definition, image geometric deviation and image rotation angle were introduced, which makes a comprehensive evaluation to dorsal hand vein images. A stepwise quality evaluation system was proposed. Different thresholds were established in different stages, and only the images that met the needs of the system are able to enter the next evaluation stage. In this way, evaluation time is reduced effectively, and the quality and velocity of selected images can be guaranteed. In quality assessment based on multiple measure fusion, a novel method to determine the optimal weights of the different parameters based on the average slope of the image quality and the recognition rate curve was proposed. And the curve refers to influence on recognition rate curve when the five dorsal hand vein image quality parameters make a change. Then weight average method was used to fuse the evaluation of each dorsal hand vein image quality parameter. Eventually the final score of image quality was acquired. And a large amount of experimental data shows that the final score of objective evaluation method presented in this article is in accordance with the subjective evaluation value and can ensure the best recognition rate.(2) Because of the trend of distributed system, heterogeneity dorsal hand vein image quality evaluation was proposed. Firstly the concept of heterogeneity image was described. Based on the principle of distributed system, three dorsal hand vein image libraries were collected. Then this paper made comprehensive analysis of heterogeneity images combined with its recognition rate. Finally, the most important parameter cause of heterogeneous was proposed. It can provide a reference for future research.