In recent years, the development of steganalysis based on convolutional neural networks (CNN) has brought new challenges to the security of image steganography. However, the current steganographic methods are difficult to resist the detection of CNN-based steganalyzers. To solve this problem, we propose an end-to-end image steganographic scheme based on generative adversarial networks (GAN) with adversarial attack and pixel-wise deep fusion. There are mainly four modules in the proposed scheme: the universal adversarial network is utilized in Attack module to fool CNN-based steganalyzers for enhancing security; Encoder module is seen as the generator to implement the pixel-wise deep fusion for imperceptible information embedding with high payload; Decoder module is responsible for the process of recovering embedded information; Critic module is designed for the discriminator to provide objective scores and conduct adversarial training. Besides, multiple loss functions together with Wasserstein GAN strategy are applied to enhance the stability and availability of the proposed scheme. Experiments on different datasets have verified the advantages of adding universal adversarial perturbations for higher security against CNN-based steganalyzers without compromising imperceptibility. Compared with state-of-the-art methods, the proposed scheme has achieved better performance in security.
相似文献This paper presents a reversible image watermarking (RIW) method including an adaptive feedback part based on difference expansion (DE). With respect to some parameters of the image, peak signal to noise ratio (PSNR), the highest payload capacity and the corresponding embedding threshold are spontaneously calculated by using the proposed adaptive feedback-based reversible Image watermarking (AFRIW). The payload capacity for data embedding is briefly explained. The machinery part of the adaptive feedback for controlling the payload capacity is presented. Software verification of three cover images is presented. Based on some image parameters, the comparative result between the proposed AFRIW algorithm and DE-based RIW method is presented. This paper also presents the VLSI architecture of this proposed algorithm for RIW. The proposed architecture has been implemented using VIVADO 2016.2 based on Xilinx Virtex-7 FPGA and Zynq device platforms. The software implementation results clearly demonstrated that the AFRIW method provides higher PSNR than the DE-based RIW method. The hardware implementation results indicate that the proposed algorithm has low timing complexity over other existing feedback based RIW algorithms which in turn provide higher speed.
相似文献This paper introduces a deep learning-based Steganography method for hiding secret information within the cover image. For this, we use a convolutional neural network (CNN) with Deep Supervision based edge detector, which can retain more edge pixels over conventional edge detection algorithms. Initially, the cover image is pre-processed by masking the last 5-bits of each pixel. The said edge detector model is then applied to obtain a gray-scale edge map. To get the prominent edge information, the gray-scale edge map is converted into a binary version using both global and adaptive binarization schemes. The purpose of using different binarization techniques is to prove the less sensitive nature of the edge detection method to the thresholding approaches. Our rule for embedding secret bits within the cover image is as follows: more bits into the edge pixels while fewer bits into the non-edge pixels. Experimental outcomes on various standard images confirm that compared to state-of-the-art methods, the proposed method achieves a higher payload.
相似文献The existed digital steganography models and theories are not effective enough to guide the steganography processing. Based on previous studies, this paper proposes a complete digital steganography model based on additive noise. And then, with security analysis from KL divergence, the embedding optimization strategy is given through theoretical derivation needless of any side information: optimizing the embedding modification position and optimizing the embedding modification direction (+1 or???1). Through theoretical derivation, we also obtain the quantitative relationship between the pixels modification probability and the adjacent pixels difference, and prove that modification by ±1 randomly cannot enhance steganographic security definitely. The research in this paper can provide theoretical guidance for the design of steganography algorithms. Compared with previous studies, the proposed embedding optimization strategy has outstanding advantages of being easy to implement and being effective to improve steganographic security. The experiments by optimizing LSBM and MG algorithms show that the proposed embedding optimization strategy can effectively improve each algorithm’s steganographic security at a relative small payload.
相似文献Social media platform such as WeChat provides rich cover images for covert communication by steganography. However, in order to save band-width, storage space and make images load faster, the images often will be compressed, which makes the image steganography algorithms designed for lossless network channels unusable. Based on DCT and SVD in nonsubsampled shearlet transform domain, a robust JPEG steganography algorithm is proposed, which can resist image compression and correctly extract the embedded secret message from the compressed stego image. First, by combining the advantages of nonsubsampled shearlet transform, DCT and SVD, the construction method for robust embedding domain is proposed. Then, based on minimal distortion principle, the framework of the proposed robust JPEG steganography algorithm is given and the key steps are described in details. The experimental results show that the proposed JPEG steganography algorithm can achieve competitive robustness and anti-detection capability in contrast to the state-of-the-art robust steganography algorithms. Moreover, it can extract the secret message correctly even if the stego image is compressed by WeChat.
相似文献Unsupervised representation learning of unlabeled multimedia data is important yet challenging problem for their indexing, clustering, and retrieval. There have been many attempts to learn representation from a collection of unlabeled 2D images. In contrast, however, less attention has been paid to unsupervised representation learning for unordered sets of high-dimensional feature vectors, which are often used to describe multimedia data. One such example is set of local visual features to describe a 2D image. This paper proposes a novel algorithm called Feature Set Aggregator (FSA) for accurate and efficient comparison among sets of high-dimensional features. FSA learns representation, or embedding, of unordered feature sets via optimization using a combination of two training objectives, that are, set reconstruction and set embedding, carefully designed for set-to-set comparison. Experimental evaluation under three multimedia information retrieval scenarios using 3D shapes, 2D images, and text documents demonstrates efficacy as well as generality of the proposed algorithm.
相似文献At present, the binary images are often used as the original watermark images of many watermarking methods, but partial methods cannot be easily extended to colour image watermarking methods. For resolving this problem, we propose a new watermarking method using ternary coding and QR decomposition for colour image. In the procedure of embedding watermark, the colour image watermark is coded to ternary information; the colour host image is also separated into image blocks of sized 3?×?3, and these image blocks are further decomposed via QR decomposition; then, one ternary watermark is embedded into one orthogonal matrix Q of QR decomposition by the proposed rules. In the procedure of extracting watermark, the proposed method uses the blind-manner to extract the embedded ternary information. The novelty of this scheme lies in the proposed ternary coding for watermark image, which can improve the imperceptibility, embedded watermark capacity and real-time feature of the watermarking scheme. The results of simulation show the presented technique is better than other compared schemes with respect to imperceptibility, embedded watermark capacity and real-time feature under the similar robustness.
相似文献Hiding sensitive information in a host image (or 2D signal) is a challenging task. Several image steganography techniques have been proposed in recent years, which either have low embedding capacity, or the embedded images are vulnerable. The proposed technique, which is based on Golden Ratio and Non-Subsampled Contourlet Transform (GRNSCT) model provides both high embedding capacity as well as the confidentiality of the embedded images. The high embedding capacity is achieved via a combination of mosaic process and two level NSCT (Non-Subsampled Contourlet Transform), while confidentiality is attained via double layer encryption based on shuffling method of a deck of cards. Several types of security evaluation metrics, such as, key sensitivity, histogram, and information entropy, are utilized to assess the robustness of the embedded images. The experimental results demonstrate that the proposed multi-image steganography technique achieves 24 bpp (bits per pixel) embedding capacity, or 300% payload with PSNR up to 42.38 dB (decibels), which is better than the existing techniques.
相似文献Nowadays, with the development of public network usage, medical information is transmitted throughout the hospitals. A watermarking system can help for the confidentiality of medical information distributed over the internet. In medical images, regions-of-interest (ROI) contain diagnostic information. The watermark should be embedded only into non-regions-of-interest (NROI) regions to keep diagnostically important details without distortion. Recently, ROI based watermarking has attracted the attention of the medical research community. The ROI map can be used as an embedding key for improving confidentiality protection purposes. However, in most existing works, the ROI map that is used for the embedding process must be sent as side-information along with the watermarked image. This side information is a disadvantage and makes the extraction process non-blind. Also, most existing algorithms do not recover NROI of the original cover image after the extraction of the watermark. In this paper, we propose a framework for blind diagnostically-lossless watermarking, which iteratively embeds only into NROI. The significance of the proposed framework is in satisfying the confidentiality of the patient information through a blind watermarking system, while it preserves diagnostic/medical information of the image throughout the watermarking process. A deep neural network is used to recognize the ROI map in the embedding, extraction, and recovery processes. In the extraction process, the same ROI map of the embedding process is recognized without requiring any additional information. Hence, the watermark is blindly extracted from the NROI. Furthermore, a three-layer fully connected neural network is used for the detection of distorted NROI blocks in the recovery process to recover the distorted NROI blocks to their original form. The proposed framework is compared with one lossless watermarking algorithm. Experimental results demonstrate the superiority of the proposed framework in terms of side information.
相似文献