Encrypted Image Encoding And Decoding Methods Based On Compressed Sensing

The encryption data compressive theory proposed by Mark Johnsonand Prakash Ishwar is different from traditional encrypted methods,which can directly code the encryption data with compressed theory, andsuccessfully reconstruct the original information after transmission. Dueto the coded data has been encrypted, the above method awfully ensurethe security of original data. Compared with other encryptedtechnologies, encryption data compressive has a promising applicationin signal and image encryption field.After analyzing the existing encrypted image coding andreconstruction methods based on signal processing，Multiple DescriptionCoding (MDC) and split Bregman iterative methods greatly improve therobustness and quality of the image reconstruction,also reduce the timerequired for image reconstruction. The main works and contributions ofthis thesis can be described as follows:(1) Considering that the sampled data transmission may bevulnerable to noise, coding errors, packet loss and other problems inharsh channel， we adopted an multiple description coding anddecoding methods of encrypted image based on compressed sensing. Weapplied CS methods and multiple description coding respectively intotransmission, then reconstruct the original image with CS algorithm andTotal Variation (TV). The results show that the proposed methods cannot only efficiently reduce the computational complexity, but alsogreatly enhance the recovery image quality.(2) Based on the Compressive Sensing and encrypted datacompression theory, a new encryption image reconstruction algorithm isproposed. Compared to traditional optimization algorithm, the proposedmethod introduces split Bregman iterative algorithm to the process,getting the measurement coefficients from the decoder. Simultaneously, based on the features of the encrypted image, we add a new constraintfunction which can improve the quality of the image reconstruction andreduce the time of the reconstruction.