| In recent years, the information security has become a hot issue. Biometrics recognition technology possesses many advantages, such as abundant information, stable feature, collecting easily,Which provides a good plan to resolve the information security. With the development of multimedia technology and networks, remote biometrics recognition will be widely employed. But biological information is tampered easily in the network, which caused huge losses for government, military, business.Information hiding is an information security technology, which can protect the multimedia information effectively. We can use the information hiding technology to protect biometric image and to achieve remote biometric recognition. The traditional information hiding method can't solve effectively to hide the biometric image. We find a certain correlation between the image and the image, which can reconstruct images mutually. In our paper, we propose a novel frequency domain embedding algorithm based on the correlation analysis , which can solve effectively to hide the secret information of large capacity.This article focuses on two parts: one is researching the theory of information hiding and describing the characteristic of information hiding method. We discuss mainly the strategy of embedding biological image and indicate the type of the attacking biological image in the transmission process. In this paper, we propose a information hiding method which is based on human visual model and correlation analysis combining. Second, we analyze the characteristics of biological images deeply and take into the information redundancy in the inter-class and between-class. Those information redundancy bring difficulties for information hiding and interfere with the biological image recognition. In this paper, we propose a new feature extraction method which is based on an adaptive covariance matrix to reduce the dimensionality of biological image. Thereby it can achieve the remote biometric image recognition. |
PDF Full Text Request