Research On Reversible Data Hiding Algorithm In Encrypted Images Based On Prediction Of Pixel Sort And Classify

With the rapid development of network and cloud computing,cloud storage gradually appears in people’s lives.Users upload encrypted images to the cloud,and cloud managers embed data such as standard tags in images for image management of cloud.Based on the need for secure storage of data in the cloud,Reversible Data Hiding in Encrypted Image(RDH-EI)technology has become a research hotspot.In this thesis,the block mod256-block scrambling encryption algorithm is used to retain the correlation between adjacent pixels,and combined with reversible data hiding algorithm which based on multi-level prediction error histogram shifting,aroundding the pixel prediction accuracy,additional data embedding rate and direct decryption image quality,this thesis proposes two completely reversible encrypted image reversible data hiding algorithms.The main research contents of this thesis are as follows:This thesis proposes the RDH-EI based on pixel sorting and classify.In the image encryption stage,the block mod256-block scrambling algorithm is used to obtain the encrypted image;in the additional data hiding stage,the encrypted image is first divided into blocks and sorted the pixels in the block with ascending order,then used the pixel classification prediction in the block to obtain the predicted value of the pixel in the block,finally the multi-level prediction error histogram shifting and the turn prediction embedding model are used to embed additional data in the encrypted image.Experimental results show that the algorithm can extract additional data without error and restore the original image without loss;the use of sorting classification prediction method effectively improves the accuracy of pixel prediction,and the additional data embedding rate can reach 0.92 bpp when EL = 4,which is better than the existing similar algorithm,while directly decrypted the image quality is better.This thesis proposes an adaptive RDH-EI algorithm based on pixel classify and sorting prediction method.Based on the idea of pixel sorting,a pixel classify sorting prediction method is designed.This method divides pixels into three types,and according to the magnitude relationship between the difference between the predicted pixel and the maximum and minimum values of the sorted pixels and the parameter EL,design different prediction methods for the predicted pixels.On this basis,an adaptive algorithm is proposed,when embedding additional data,an turn embedding model is used to embed additional data in three types of pixels.The experimental results show that using pixel classify sorting prediction improved the prediction error ratio,so that the adaptive algorithm has a larger range of additional data embedding,and the quality of the direct decryption image is also better than other algorithms.Finally,in order to verify the feasibility and effectiveness of the proposed algorithm,the encrypted image reversible data hiding algorithm based on pixel sort classify prediction is selected,a GUI simulation system is designed to demonstrate and verify the algorithm intuitively.