Capacity Research And Analysis Of Dorsal Hand Vein Image

Following the development of IOT and Big Data, biometrics identification has been moving towards large-scale era gradually. Biometric feature capacity research is the fundament to given a biometric system for large-scale personal identification. Biometric capacity is defined as the maximum number of individuals under an acceptable system identification rate.As a kind of new rising biometrics image, dorsal hand vein image attract more and more researcher’s attention. Most of researches focus on feature extraction method or identification method. But very little works pay attention to the potential capacity of dorsal hand vein image. In this paper, we proposed a capacity estimation method for dorsal hand vein image which based on block transform coding, and verified it through some contrast experiments. The detail work in this paper as follow:(1) We compare the entropy of few biometric feature images to verify the feasibility of hand vein image firstly.(2) More work focus on its capacity estimation. It is addressed by drawing parallels between compression and classification of images, and it leads to an empirical approach for capacity estimation based on adaptation of block transform coding that is normally used for image compression. The proposed method is shown to consist of image transform to texture based feature space, quantization and binary encoding of statistical texture descriptor, followed by capacity estimation based on the statistical variability among the resulting binary codes of a given biometric database. Applying proposed method, the biometric capacity of dorsal hand vein images is estimated for the first time to be approximately 105 for an area of 2268 mm2.(3) In order to verify the proposed method, we also use it in fingerprint and face, and compute these capacities are few millions and few hundred millions that agree on reference or some other reports. It means that our method is correct.