Embedded Watermark Information Research Articles

AI-assisted deepfake detection using adaptive blind image watermarking

This paper proposes a new adaptive blind watermarking technology for deepfake detection, which can embed deepfake detection information into the image and verify the image's authenticity without requiring additional information. The proposed scheme utilizes mixed modulation combined with partly sign-altered mean value to embed a set of coefficients that enhance robustness against attacks while maintaining high image quality. Additionally, blind adaptive deepfake detection technology with the tamper detection mean value is employed to detect relative positions adaptively, even when face images are slightly modified or deepfaked. To further improve the performance of the proposed scheme, a gray wolf optimizer is introduced to optimize parameters, and a denoising autoencoder is employed to facilitate the identification of extracted watermarks. This technology will adaptively embed watermark information while preserving the original face image, thereby maintaining the authenticity of the face in the image and verifying the owner of the image. The code is available at https://github.com/lyhsu01/AwDD.

A robust blind color watermarking algorithm based on the Radon-DCT transform

Digital watermarking is an effective technique for image copyright protection. Many digital image watermarking algorithms are sensitive to geometric distortions. These distortions make it difficult to detect and extract watermarks. In this study, a robust color watermarking algorithm combining the Radon transform and DCT transform is proposed. First, the color carrier image is converted to the YUV color space from RGB, and the U component of YUV is transformed to the Radon domain using the Radon transform. Then, a two-dimensional 8 × 8 discrete cosine transform (2D-DCT) is carried out on the selected blocks of the Radon domain, and some fixed midfrequency coefficients are selected to embed watermark information. Moreover, the Arnold transform is applied to encrypt the watermarks and the random permutation function to scramble the embedding positions. Finally, the watermark embedding and blind extraction processes are completed by modifying the midfrequency coefficients using the presented rules. A series of simulation experiments show that the watermark algorithm has high imperceptibility and good robustness to various attacks.

Visible watermarking for 3D models based on 3D Boolean operation

With the continuous expansion of the application field of 3D models and the convenience of model storage, transmission, and replication, the problems of illegal reproduction, illegal modification, and embezzlement faced by 3D models are gradually becoming severe, which has brought substantial economic losses to the producers of 3D models. Therefore, this study aims to develop a visible watermarking method for 3D model using 3D Boolean operation. First, based on TrueType font library, the 3D model of the watermark is generated through the extraction, positioning and combination of character outline. Next, the spatial registration between the watermark model and the 3D model was carried out according to the embedding position of the watermark. Finally, based on multiple 3D Boolean operations, the copyright information of the 3D watermark model is embedded into the 3D model in various modes, such as embossing, debossing, and surface mesh embedding. The experiments demonstrate that the method of this paper supports embedding watermark information in B-rep, CSG, and voxel models. This method can readily support business applications for visible watermarking of 3D models and has laid a good foundation for researching multipurpose 3D model watermarking algorithms supporting copyright notification and protection.

An efficient hexadecimal network flow watermark method for tracking attack traffic

Network flow watermark technology is a traffic marking technique that embeds watermark information into the characteristics of network flows to mark and trace attack flows generated by network attackers. However, with the development of network attack techniques, the time and number of packets required for network attacks have decreased. Existing network flow watermark technologies fail to balance watermark robustness and efficiency, resulting in poor practicality. To address this issue, this paper proposes an efficient hexadecimal network flow watermark method. The method introduces an efficient interval watermark algorithm and utilizes an interval synchronization algorithm to self-learn watermark parameters, thereby improving the encoding efficiency of the watermark. The design of watermark start and end markers ensures the practicality of network watermarks, enabling traceability and source attribution of attack flows in real network environments. The proposed method is experimentally tested using real network traffic, and the results demonstrate that even in the presence of a network jitter, the watermark detection success rate of this scheme remains above 95%. Compared to other network flow watermark schemes, the hexadecimal network flow watermark proposed in this paper achieves a 50% improvement in encoding and decoding efficiency while ensuring robustness. It also exhibits excellent resistance to network jitter, packet loss, and false packet insertion.

Three-dimensional optical watermarking scheme using compressed sensing in integral imaging system

The elemental image array (EIA) is a special form of the integral imaging three-dimensional (3D) display system for recording, transmitting and displaying 3D image information, which consists of a series of small image units taken from different perspectives. There is an urgent need for protecting the copyright of EIA to improve security and anti-aggressive capability. In this paper, we propose a 3D optical watermarking scheme for integral imaging system based on compressed sensing (CS) theory. The proposed watermarking scheme consists of two subsystems, one is the EIA watermark compression and embedding subsystem, and the other is the EIA watermark extraction and display subsystem. CS theory using by phase coding mask is applied for realizing the EIA encryption and compression simultaneously, which can reduce the total amount of embedded watermark information, expand the space of secret keys and make the system more flexible and compact. The experimental results are presented to demonstrate the feasibility and effectiveness of the proposed method. The parameters of the integral imaging watermark encryption system, including the focal length and aperture of the lenslet array, the object distance, the wavelength, the diffraction distance, and the measurement matrix of CS, can all be used as the secret keys. The method proposed in this paper can effectively broaden the keys space and enhance the security of the integral image 3D display system.

Research on Robust Digital Watermarking Based on Reversible Information Hiding

The development of modern Internet communication technology and the popularization of multimedia technology have brought convenience to the sharing and storage of multimedia information such as images, videos, and audio. However, at the same time, it has brought about the problem of copyright theft of multimedia information, causing serious information security risks. Digital watermarking technology embeds copyright information in multimedia information in an invisible way, which can effectively realize copyright protection and traceability of infringement. Aiming at the problem that the existing learning model-based methods cannot fully extract and fuse the features of carrier images and watermark images, a robust digital watermarking method based on reversible information hiding is proposed. First, a watermark embedding model based on reversible information hiding is established, and the features of the download volume image and the watermark image in different dimensions are fully extracted and fused to generate a dense image with excellent visual quality. Then, a watermark extraction model based on reversible information hiding is established, and a noise layer is added between the embedding and the extraction model, and the attacked dense image is input to the watermark extraction model to extract the watermark. Under the constraint of the loss function, the network model learns to embed watermark information in the area that is more robust to the attack and is not easy to cause visual quality degradation, so as to optimize the comprehensive performance of the method. Experimental results show that the proposed method effectively improves the imperceptibility and robustness.

Anti-pruning multi-watermarking for ownership proof of steganographic autoencoders

Model watermarking Model watermarking is a method for embedding watermark information into a neural network model. It proves the ownership of the model without affecting its performance. Since there are plenty of attacks against model pruning, it becomes more significant to design anti-pruning model watermarking algorithms. In this paper, multiple watermark embedding is performed to protect the model copyright for the image steganography auto-encoder model “Hiding Data with Deep Networks” (HiDDeN). Firstly, the appropriate model weights are selected by employing three classical model pruning algorithms of model weights. Secondly, the model watermark is spread by using Discrete Cosine Transform (DCT)-based image watermarking algorithm, which improves the noise and pruning resistance of the model watermark. Finally, the model watermark is embedded to the 4th and 5th decimal places of the selected model weights. The experimental results demonstrate that the proposed algorithm has a good robustness against model pruning without affecting the watermark extraction performance of the auto-encoder network model. Even with the embedded model watermark, the decoder's watermark extraction accuracy is still higher than 0.9993. and the autoencoder is still valuable when the model weights are pruned by 40%. Furthermore, the proposed algorithm has a certain degree of improvements in watermarking capacity.

An Effective Framework for Intellectual Property Protection of NLG Models

Natural language generation (NLG) models combined with increasingly mature and powerful deep learning techniques have been widely used in recent years. Deployed NLG models in practical applications may be stolen or used illegally, and watermarking has become an important tool to protect the Intellectual Property (IP) of these deep models. Watermarking technique designs algorithms to embed watermark information and extracts watermark information for IP identification of NLG models can be seen as a symmetric signal processing problem. In terms of IP protection of NLG models, however, the existing watermarking approaches cannot provide reliable and timely model protection and prevent illegal users from utilizing the original performance of the stolen models. In addition, the quality of watermarked text sequences generated by some watermarking approaches is not high. In view of these, this paper proposes two embedding schemes to the hidden memory state of the RNN to protect the IP of NLG models for different tasks. Besides, we add a language model loss to the model decoder to improve the grammatical correctness of the output text sequences. During the experiments, it is proved that our approach does not compromise the performance of the original NLG models on the corresponding datasets and outputs high-quality text sequences, while forged secret keys will generate unusable NLG models, thus defeating the purpose of model infringement. Besides, we also conduct sufficient experiments to prove that the proposed model has strong robustness under different attacks.

A novel watermarking scheme for medical image using support vector machine and lifting wavelet transform.

Digital image watermarking has become a valuable tool for preventing unauthorized use and alteration of digital images due to technological advancements. A fundamental difficulty in digital image watermarking is to provide resilience against geometrical assault while retaining a sufficient degree of imperceptibility and security. This study presents an efficient authentication scheme for digital image watermarking on medical images benefiting from both techniques: Support Vector Machine (SVM) and Lifting Wavelet Transform (LWT). In this article, we use two strategies, where SVM is used first to separate the Region of Interest (ROI) from the Non-Region of Interest (NROI) in the medical image. Then LWT is applied to embed watermark information within the NROI part of the medical image (Cover Image). Moreover, we have applied a shared secret key to enhancing the robustness of the proposed scheme. The method is tested on an extensive image database to see how it performs under different situations. The research looked into the various experimental analyses to establish the acceptability of the existing scheme. The simulation is performed to measure the imperceptibility and robustness using various evaluation metrics.

Geometrical attacks resilient statistical watermark decoder using polar harmonic Fourier moments

This paper presents a new robust multiplicative watermark detector. Due to the strong robustness against various attacks, polar harmonic Fourier moment (PHFM) magnitudes are used as the employed watermark carrier. The distribution of PHFM magnitudes is highly non-Gaussian and can be properly modeled by a heavy-tailed probability density function (PDF). In this paper, we proved that Weibull distribution can suitably fit the distribution of PHFM magnitudes, and based on this, we presented a statistics-based watermark decoder by using the Weibull as a prior for the PHFM magnitudes. In watermark embedding, a multiplicative manner was used to embed watermark information in PHFM magnitudes of the highest entropy blocks to achieve better robustness and imperceptibility. In watermark detection, we developed a Weibull distribution-based statistical watermark decoder, which uses the maximum likelihood (ML) decision rule. Compared with Bessel K form (BKF), Cauchy, and generalized Gaussian (GG)-based decoders, the Weibull-based decoder demonstrates stronger robustness. In addition, the proposed watermark decoder is more robust against geometrical and common image processing attacks than existing statistical watermark decoders.

A robust and secure watermarking algorithm based on DWT and SVD in the fractional order fourier transform domain

The present paper aims to develop and validate a robust and secure watermarking algorithm. Firstly, the algorithm transforms the cover image and the watermark image separately by FRFT to obtain the amplitudes of both the cover image and watermark image, and then the two-level DWT transformation only carries out for the amplitude of the cover image. Secondly, the algorithm applies the SVD to the low-frequency sub-band of the second level DWT of the cover image and the amplitude of the watermark image. Thirdly, a new matrix is constructed based on the singular values from both the cover and watermark images to embed watermark information. Lastly, the algorithm takes the preliminary numerical calculation to determine an appropriate transformation order of FRFT for the cover image and to provide an outstanding balance between the imperceptibility and robustness in the proposed watermarking scheme. Moreover, it guarantees the security of watermarking scheme by the transformation order of FRFT. Experimental results demonstrate that the proposed watermarking scheme provides better performance in imperceptibility and resistance against traditional signal processing and geometric attacks, especially in image rotation, image cropping, average filtering, median filtering, and Gaussian filtering in comparison with the existing methods in the literature.

A blind color image watermarking algorithm combined spatial domain and SVD

It has great significance for protecting the image copyright with color image digital watermarking technique, which needs the watermarking algorithm has high robustness, good imperceptibility and short running-time. A blind colour image watermarking algorithm combined spatial domain and singular value decomposition (SVD) is designed to protect the rightful ownership of color image in this paper. The theoretical principle combined SVD and spatial domain is first discussed and proposed in this paper. Based on the derived principle, the watermark information can be directly embedded into the maximum singular value (MSV) of the host image in the spatial domain by the new quantization index modulation with variable quantization step, and the embedded watermark information can be directly extracted from the MSV of the watermarked image block in the spatial domain with geometrical correction. The contributions of this paper include the following four points: (1) the relationship between the changed MSV of SVD and the average changed value of image pixel in the spatial domain is the first derived in this paper; (2) the presented method has the merits of both spatial-domain technique and transform-domain technique; (3) the relationships of each component of the color image are used to improve the invisibility and the geometrical correction of watermarked image is also adapted to further improve the robustness; and (4) this presented method overcomes the false-positive problem that existed in many watermarking methods based on SVD. Simulation results demonstrate the new presented color image watermarking technique has better imperceptibility, higher robustness and shorter running-time compared with some state-of-the-art methods.

Robust Blind Watermarking Algorithm Based on Contourlet Transform with Singular Value Decomposition

In this Letter, we propose a blind and robust multiple watermarking scheme using Contourlet transform and singular value decomposition (SVD). The host image is first decomposed by Contourlet transform. Singular values of Contourlet coefficient blocks are adopted to embed watermark information, and a fast calculation method is proposed to avoid the heavy computation of SVD. The watermark is embedded in both low and high frequency Contourlet coefficients to increase the robustness against various attacks. Moreover, the proposed scheme intrinsically exploits the characteristics of human visual system and thus can ensure the invisibility of the watermark. Simulation results show that the proposed scheme outperforms other related methods in terms of both robustness and execution time.

Large-capacity reversible image watermarking based on improved DE.

Aiming at solving the problems of bad imperceptibility and low embedding rate of existing algorithms, a novel large-capacity reversible image watermarking based on improved difference expansion (DE) is proposed. Firstly, the smoothness calculation algorithm is used to calculate and sort the smoothness values of the divided image sub-blocks; then, the scrambled watermark is embedded into the sub-blocks with less smoothness after removing the abrupt point by using the generalized difference expansion (GDE); finally, the absolute difference operation is applied to the generated overflow pixels to make their pixel values within a reasonable range for embedding watermark information. Under the premise of ensuring a certain visual quality, multiple watermark embedding can effectively improve the embedding rate. The simulation results show that this algorithm not only realizes blind extraction, but also recovers the original images without loss. At the same time, this algorithm achieves a high embedding rate (the average embedding rate is as high as 77.91 dB) without decreasing the visual quality.

A Blind Watermarking Algorithm for Digital Image Based on DWT

In order to improve the robustness of digital image watermarking algorithms, the digital image blind watermarking algorithm proposed in this paper embeds watermark information in a reasonable position. First, scramble the watermark information and embed the random signal into the low-frequency domain of the image. On the one hand, it can effectively prevent attacks on the watermark system and enhance the security of the system, disperse as much as possible. In the watermark embedding process, a method based on the wavelet transform domain is used. The original image is not required to participate in the watermark extraction process, and the blind watermark function is implemented. In order to verify the effectiveness of the algorithm, the watermarked image is subjected to filtering, scaling, noise, cropping and rotation attacks, and the peak signal-to-noise ratio and normalized correlation coefficient are used to quantitatively evaluate the watermarked image. Experimental results show that the algorithm can resist a variety of attacks, and has good imperceptibility and robustness.

Vulnerability Attacks of SVD-Based Video Watermarking Scheme in an IoT Environment

This study reviews the superiority and limitations of the Singular Value Decomposition (SVD)-based video watermarking scheme. First applies key frame selection to the video sequence to obtain an appropriate video frame for the purpose of watermark embedding in this paper. The Fibonacci-Lucas image transformation scrambles the grayscale watermark image before the embedding process. The former scheme satisfies the imperceptibility and robustness aspects. However, it can suffer from the False-Positive-Problem (FPP) in which a counterfeit watermark image can be easily reconstructed by a malicious attacker. This study presents a new technique for overcoming the FPP of the aforementioned scheme. The proposed method embeds the principal components of the watermark image rather than the Singular Value Matrix (SVM) into the host video. As results, the proposed methods yield promising results in terms of imperceptibility as well as robustness. At the same time, it solves the FPP. A theoretical analysis of embedded watermark information is also provided in this work. The theoretical analysis meets the experimental-based results in the SVD-based video watermarking scheme.

A High Capacity Watermarking Technique for the Printed Document

Digital watermarking technology is an effective method for copyright protection of digital information, such as images, documents, etc. In this paper, we propose a high capacity text image watermarking technique against printing and scanning processes. Firstly, this method obtains the invariant in the process of printing and scanning under the mathematical hypothesis model of print-scan transformation. Then based on the print-scan invariant, the Fourier descriptor is used to flip the trivial pixel points with high frequency information on the character boundary. Next, considering the resolution of the print-scan equipment and its influence on the print-scan invariant, a quadratic quantization function is proposed to embed watermark information of multiple bits for a single character. Finally, the QR code (Quick Response code) is researched, which has large information capacity, robust error correction ability and high decoding reliability. By using the QR code as the watermark information, we can reduce the impact of bit error rate during watermark extraction, and the robustness of the watermark information can be improved. The experimental results show that the proposed text watermarking algorithm has the advantages of anti-print scanning, anti-scaling, large capacity and good visual effects.

Anti-HEVC Recompression Video Watermarking Algorithm Based on the All Phase Biorthogonal Transform and SVD

ABSTRACTWith the rapid development of Internet of Things and multimedia technologies, people are paying more attention to the security of images and videos. To further improve the security of video content, a video watermarking algorithm based on the high efficiency video coding (HEVC) compression domain is proposed in this paper. The scheme embeds watermark information in the quantized coefficients, which can solve the cumulative error caused by watermark embedding. Moreover, the all phase biorthogonal transform (APBT), an improved transform, is used to compress the watermark data at the digital watermark preprocessing stage, optimize the watermark embedding scheme, and improve the embedding capacity of video watermarking. Due to the characteristics of HEVC video coding, most HEVC video watermarking schemes based on compression domain cannot resist recompression attack. To solve this problem, on the basis of APBT and singular value decomposition, an anti-HEVC recompression video watermarking algorithm is proposed and can effectively resist the HEVC recompression. Experimental results show that even when the quantization parameter is larger than 40, the extracted watermark still has good image quality. In addition, the algorithm also has a good performance in noise resistance.

An improved color image watermarking scheme based on Schur decomposition

A novel colour image watermarking scheme based on Schur decomposition is proposed in this paper. By analysing the 3 × 3 unitary matrix U via Schur decomposition, the strong similarity relation between U2,1 and U3,1 is found. This relation will represent one kind of embedded watermark information, i.e., '1' or '0', and will be used to extract watermark from 3 × 3 unitary matrix U in the blind manner. Moreover, the proposed method requires lower computational complexity than other methods based on singular value decomposition (SVD) since Schur decomposition is only an intermediate step in SVD decomposition. In addition, the proposed method does not have the false-positive detection problem like that based on SVD. The experimental results show that the proposed scheme has better performance in terms of invisibility, robustness, capacity and real-time feature.

Visual perception based robust watermarking with integral imaging

This paper presents a robust watermarking approach for hiding gray watermark into digital images. Conventional transform-based watermarking schemes provide only one transform plane for watermark embedding, which makes easy to remove embedded watermark information. Cellular automata transform (CAT) based watermarking scheme can address this problem with CA properties. In this paper, the watermark is first recorded in the form of an elemental image array (EIA) with the computational integral imaging (CII) system, and then the recorded EIA is embedded into the CAT coefficient. The watermark embedding coefficient will be determined by an alpha map which generated by saliency map. This proposed method can provide a more robust watermarking system because EIA is composed of many elemental images and each image has its own property of the watermark. Even though most data of elemental images are lost, the watermark can be reconstructed successfully by using the remaining elemental images. Furthermore, the alpha map which is generated by saliency map can make watermark more invisible for the human visual system. Some experimental results show competence of our system.