Research On High Efficiency Image Security Coding Algorithm For Biometric Images

The rapid development of Internet technologies bring the age of information sharing and exchanging.At the same time,with the deeper research and development of fingerprint,palm-print,iris and face recognition technology,biometric recognition are widely used in the fields such as military area,biomedical and political affair which require high performance of information security because of its universality and uniqueness of biometric.Biometric images are stored as the digital image in a digital format,thus it cost large storage space.Kinds of data related to biometric images might be broken authority and stolen passwords,during the transmission,because of limited transmission bandwidth and storage space.Therefore,we combined image compression with image encryption.These two technologies could ensure the decrypted image without distortion while decreasing the storage space of cipher-text and realize the encryption of biometric images effectively,thereby protecting the security of the information.We achieve image compression while implementing the process of image encryption in order to satisfy the high-speed and efficient requirement of modern science technology.In this paper,we mainly research into the process of compressive sensing and achieve the research of biometric image encryption algorithm which combine with compressive sensing to make the algorithm achieve encryption effect and biometric information protection.And we make sure the image decryption and reconstruction not affect the accuracy of the biometric recognition technology and simulation results demonstrate good quality of the reconstruction.In this paper,we mainly carry out the following two aspects of the work:(1)We propose a biometric image encryption algorithm based on Dual-tree Complex Wavelet Transform(DT-CWT)and Compressive Sensing(CS).We adopt Dual-tree Complex Wavelet Transform to sparsely represent the biometric image,it can overcome Discrete Wavelet Transform's defects,such as translation sensitivity and lack of direction sensitivity.Also it can enhance the image edge and texture information.Because of this we can attain more sparse transformation coefficients.In order to reduce the number of nonzero coefficients and data redundancy,we employ the noise shaping gather maximum energy of DT-CWT high frequency coefficients.In accordance with the results of computer simulation and experiment,we find that the quality of the biometric image reconstruction based on DT-CWT was higher than the method based on Discrete Wavelet Transform(DWT)and the noise shaping algorithm can improve the effects of the biometric image quality.The Fibonacci scrambling time can be used as the secrete key to guarantee the safety of encryption algorithm and make cipher-text distribution more uniform.(2)The noise has a certain effect on the quality of the image decryption and reconstruction.In order to solve this problem we integrate the robustness of the double random phase encoding(DRPE)into CS encryption algorithm.Initial condition sensitivity is the key characteristic of chaos,we select chaotic system to produce measurement matrix in order to enlarge the key space and improve the security of encryption algorithm.Through the simulation,we find that cipher-text was output as steady white noise and the gray histogram of cipher-text present Gaussian-like distribution.Thus the encryption algorithm can resist the statistical cipher-text attack and the random noising attack.In the process of encryption,the seed is used to generate random phase masks as two keys and the initial values of Logistic chaotic system are the two other keys,those could improve the security of encryption algorithm and enlarge the bigger key space.To verify the applicability of the encryption algorithm,we select standard testing images and compared experimental results with some existing academic achievements,the experimental result demonstrated that the algorithm have high security and privacy.