Research On Image Information Hiding Technology Based On Histogram Transformation

Information hiding,which hides the certain secret information in other public media,is to realize the transmission of secret information through the transmission of public media.It is often used in confidential communication and the protection of digital intellectual property rights.Image histogram is a bar graph that represents the distribution of pixels by a series of line segments or vertical stripes of unequal height.Because its gray value is not changed by the position of the carrier pixel,it is widely used in information hiding technology.However,the traditional information hiding technology based on histogram transformation has the following three problems:1.the histogram watermark embedding procedure may cause the forced flipping of the two Bin before and after the original histogram,resulting in reduced watermark image quality;2.watermark embedding rules are only simply comparing the high and low ratios of adjacent two Bin before and after,ignoring the approximate equality,resulting in reduced embedding capacity;3.during the embedding process of joint photographic experts group(JPEG)encoding format image data,the histogram of the discrete cosine transform(DCT)coefficients invalid shifting may cause the file size of the marked image to increase disproportionately.Therefore,the information hiding technology based on histogram transformation is still a research hotspot in computer science.The main work and innovation of this dissertation is as follows:(1)Aiming at the problem of reduced peak signal-to-noise ratio(PSNR)of marked images,a ternary image watermarking algorithm using two continuous Bin histogram modification is proposed.First,select the histogram embedding interval that meets the requirements of the number of pixels through an appropriate intercept threshold;then,adjust the embedding threshold according to the capacity of the watermark information to be embedded and the size of the carrier image;finally,in the case of maintaining the similar height and shape of the two continuous Bin,determine the number of modified pixels and block strategy to realize the embedding and extraction of watermark information.Due to the introduction of ternary watermark information and following the size relationship between two adjacent Bin of the original histogram,the proposed algorithm not only improves the watermark embedding capacity but also resists traditional geometric attacks compared with the existing the-state-of-art works,such as cropping and rotation,and signal processing attacks.More importantly,under the same embedding capacity,the PSNR value of image quality after watermark embedding is at least 5dB higher than the similar algorithms.(2)Aiming at the problem of low embedding capacity in the process of watermark embedding,the concept of multi-Bin and multi-ary watermark is proposed theoretically for the first time,and a generalized multi-Bin and m-ary histogram-based image watermarking algorithm is designed.First,select the appropriate histogram embedding interval;secondly,according to the watermark information to be embedded,consider the approximate equal proportions of adjacent histograms,calculate the embedding thresholds corresponding to the watermark information with equal proportions and watermark information with unequal proportions;then,according to the corresponding embedding threshold,determine the number of pixels that need to be modified;finally,the watermark information is embedded by modifying the number of pixels contained in the histogram to obtain a watermarked image.Experimental analysis shows that this method improves the watermark embedding capacity by at least 2 times compared with the existing algorithms under the premise of satisfying visual imperceptibility.(3)Aiming at the disproportionate increase in the storage size of JPEG tagged images,a reversible data hiding(RDH)algorithm for JPEG images based on adaptive location selection is proposed.This method considers not only the visual quality and embedding capacity but also the file storage size of the marked image after embedding secret information.First,divide the original JPEG carrier image into blocks,and conduct DCT transformation and quantization on each divided block to obtain the quantized alternating current(AC)coefficient;secondly,obtain the entropy value of each block according to the entropy threshold,and then divide the smooth or texture block,and arrange the divided smooth blocks in ascending order;again,in the smooth block,calculate the embedded distortion cost of each of the 63 AC coefficient positions,and arrange them in ascending order;then,the length information of secret information,entropy threshold information and position mapping information are embedded in the least significant bit(LSB)of the direct current(DC)coefficient;finally,the secret information to be embedded is sequentially embedded into the AC coefficients with the least distortion cost as '1' and '-1',get the marked image.Experimental results show that this method can perfectly recover the carrier image while extracting secret information.It is superior to the current typical algorithms.It not only guarantees the embedding capacity and obtains good visual quality,but also reduces the storage overhead in terms of file storage size.In summary,the image information hiding scheme based on histogram transformation designed in this topic can not only effectively transmit secret information or copyright information but also increase the capacity and quality of the carrier image,and the carrier image can be perfectly restored on JPEG image.This subject has practical significance.