Research On Efficient Steganographic Algorithms Based On Image Content Features

In recent years,the research on steganography with an image as a carrier is more mature,which its contribution to steganography is outstanding.The design of image steganography is mainly focusing on three indicators: perception,security and embedding capacity.However,there is a trade off between perception,security and embedding capacity,improving embedding capacity for steganography in spatial domain will lead to increase the pixel's modification amplitude or modify more pixels as the cost of distortion.Therefore,how to balance the relationship among the three indicators is the key to design an efficient steganographic algorithm.This thesis promotes the research work in which mainly focuses on the three significant indicators and embedding time of the steganography.At last,the main innovative contributions of this thesis are summarized as follows:1.Aiming at the security of adaptive pixel pair matching(APPM),the adaptive steganography in the spatial domain based on quadtree segmentation is proposed.Firstly,the consistency measure and discriminant criterion of quadtree segmentation are defined,and the optimal segmentation threshold can be selected.According to the principle of quadtree segmentation,the small image block belongs to the complex region and the large image block belongs to the smooth region.Finally,secret messages in larger n-ary are priority embeded into small image blocks to enhance the cover image quality,and secret messages in lower n-ary are embeded into big image blocks to improve the steganographic capacity,which using large embedding capacity and high security features of adaptive pixel pair matching.Experimental results show that the scheme has advantages in terms of security and embedding capacity.2.Aiming at the security of the steganographic algorithm,the embedding pattern based steganographic algorithm with embedding priority given to image textures is proposed.Combining encryption with steganography to improve the safety of the secret messages.According to the defined texture criteria and embedding capacity,the embedding regions that more difficult to be predicted and found are priority selected.Experimental results show that the proposed algorithm has advantages in resisting common steganalysis,the sensitive regions of the image are well protected.In addition,secret messages are disorderly embedded according to the embedding pattern,which further improves the security of the secret messages.3.Aiming at the security of EPES algorithm,the adaptive color image steganography in the spatial domain based on texture complexity is proposed.First,this scheme treats a channel of the color image as a reference channel.According to the texture criterion and the embedding capacity,the embedding regions of the reference channel are adaptively selected.Second,the remaining two channels of the color image search the corresponding embedding regions according to the reference channel.Finally,the three equal parts of secret messages are embeded in embedding regions of the channel respectively.The consistency of the texture sequences in the reference channel is guaranteed,the embedding mechanism simulate EPES algorithm which has high embedding efficiency and has been easily realized.Compared with traditional EPES algorithm,the security of the proposed scheme has been dramatically improved.4.Aiming at the payload of the steganographic algorithm,the edge detection based efficient steganographic algorithm with high payload is proposed.Extracting edge and non-edge regions by using most significant bits of cover pixels,and least significant bits of cover pixels are used to replace by using OPAP algorithm.Edge pixels are replaced with more LSB,and non-edge pixels are replaced with less LSB.The scheme ensures the consistency of the edge embedded sequences and edge extracted sequences,the imperception and high embedding capacity.Experimental results show that the time cost is guaranteed in 16-26 seconds,and the PSNR value is above 38 dB under the highest embedding capacity,which can quickly achieve high steganographic capacity.