Research On Reversible Information Hiding Algorithm For JPEG Images

Under the current open Internet,multimedia technology has developed rapidly and become the main particpant in the digital information age.People use digital multimedia for data communication and information exchange,which means that there is a third party to access all types of multimedia information,which will also lead to many security problems.Therefore,we are in urgent need of effective technology to enhance the security of Internet information.For the above problems,it has traditionally been alleviated by encryption and information hiding technologies.Due to the presence of garbled characters,encryption technology can easily attract attention and crack of the interceptor.The traditional information hiding technology often leads to the irreversible distortion of the carrier,which means that the carrier can not be recovered after the secret information is extracted.Unlike the above two technologies,reversible information hiding technology can not only extract secret data,but also restore and reconstruct the original carrier image.Reversible information hiding of JPEG images has far-ranging applications such as image authentication and archive management.However,due to JPEG compression usually reduces the image quality by eliminating some high-frequency information of the image,the redundancy information of JPEG image is much less than that of the uncompressed image,which also leads to the difficulty of obtaining high embedding capacity on information hiding on JPEG image.At the same time,after embedding the secret information,the file size of JPEG image will increase greatly,which will lead to the problems that the secret information is easily perceived and the image storage space is significantly increased.As for above problems,this paper mainly achieved the following research results:1.A reversible hiding algorithm for modifying quantified DCT coefficients by histogram is proposed.First,the distribution and quantity of DCT coefficients are analyzed by histogram.Then,the correlations among the number of DCT coefficients,quantization techniques and coding techniques are analyzed.Final,by analyzing the influence of modifying DCT coefficient on the visual quality and the file size of JP-EG image,we divide DCT coefficient into three classes,0,1 and 2.The algorithm applies three kinds of DCT coefficients to different ways.Under the certain embedding capacity,it effectively achieves better visual quality and suppresses the increase of image file size.2.An improved algorithm of reversible algorithm above is given.First,the frequency sensitivity and texture sensitivity of the human eye are analyzed.Then,the quantitative relationship between the three kinds of AC coefficients in the 8*8 sub blocks is statistically analyzed.It obtains that the influence of coefficients value and texture degree on histogram algorithm,and the invalid translation of "2" class coefficients will lead to the decline of image quality and the increase of image file size.Therefore,the improved algorithm is further enhance the visual quality of the image and suppress the size of the image after embedding information,by defining a block priority formula to select the suitable sub-block for embedding secret information.Last,performance analysis is carried out for the improved algorithm,such as algorithm security,complexity and reversibility,as well as the analysis of embedding capacity and change of file size of JPEG image.Through contrast experiments,we test the images with different texture levels,and then evaluate performance of the improved algorithm in terms of image visual quality,the change of file size of JPEG image and the feasibility of the algorithm.The advantages and disadvantages of the algorithm are summarized,and the problems that need to be solved and studied in the next step are given.