Watermarking Framework Research Articles

Did You Train on My Dataset? Towards Public Dataset Protection with CleanLabel Backdoor Watermarking

The huge supporting training data on the Internet has been a key factor in the success of deep learning models. However, this abundance of public-available data also raises concerns about the unauthorized exploitation of datasets for commercial purposes, which is forbidden by dataset licenses. In this paper, we propose a backdoor-based watermarking approach that serves as a general framework for safeguarding publicavailable data. By inserting a small number of watermarking samples into the dataset, our approach enables the learning model to implicitly learn a secret function set by defenders. This hidden function can then be used as a watermark to track down third-party models that use the dataset illegally. Unfortunately, existing backdoor insertion methods often entail adding arbitrary and mislabeled data to the training set, leading to a significant drop in performance and easy detection by anomaly detection algorithms. To overcome this challenge, we introduce a clean-label backdoor watermarking framework that uses imperceptible perturbations to replace mislabeled samples. As a result, the watermarking samples remain consistent with the original labels, making them difficult to detect. Our experiments on text, image, and audio datasets demonstrate that the proposed framework effectively safeguards datasets with minimal impact on original task performance. We also show that adding just 1% of watermarking samples can inject a traceable watermarking function and that our watermarking samples are stealthy and look benign upon visual inspection.

A novel gray-scale image watermarking framework using harmony search algorithm optimization of multiple scaling factors

A novel gray-scale image watermarking framework using harmony search algorithm optimization of multiple scaling factors

Security Analysis and Improvement of Dual Watermarking Framework for Multimedia Privacy Protection and Content Authentication

The demand for using multimedia network infrastructure for transmission grows with each passing day. Research scholars continue to develop new algorithms to strengthen the existing network security framework in order to ensure the privacy protection and content authentication of multimedia content and avoid causing huge economic losses. A new technology for multimedia image copyright protection and content authentication has been proposed. The innovations lie in the use of an inter-block coefficient difference algorithm to embed robust watermarking in the transform domain, and the same fragile watermark is embedded twice in the spatial domain so that any tiny tampering can be identified and located. A new encryption algorithm combined with Arnold transform is used to encrypt data before embedding. However, some security vulnerabilities were found, and successful cryptanalysis and attack were conducted. Subsequently, an improved scheme was proposed to improve the security and tamper detection ability of the original watermarking scheme and recover the tampered robust watermark. The results show that the improved scheme is safer and more reliable and shows good performance in tampering detection and the recovery robustness of the watermark.

Machine learning based multipurpose medical image watermarking.

Digital data security has become an exigent area of research due to a huge amount of data availability at present time. Some of the fields like medical imaging and medical data sharing over communication platforms require high security against counterfeit access, manipulation and other processing operations. It is essential because the changed/manipulated data may lead to erroneous judgment by medical experts and can negatively influence the human's heath. This work offers a blind and robust medical image watermarking framework using deep neural network to provide effective security solutions for medical images. During watermarking, the region of interest (ROI) data of the original image is preserved by employing the LZW (Lampel-Ziv-Welch) compression algorithm. Subsequently the robust watermark is inserted into the original image using IWT (integer wavelet transform) based embedding approach. Next, the SHA-256 algorithm-based hash keys are generated for ROI and RONI (region of non-interest) regions. The fragile watermark is then prepared by ROI recovery data and the hash keys. Further, the LSB replacement-based insertion mechanism is utilized to embed the fragile watermark into RONI embedding region of robust watermarked image. A deep neural network-based framework is used to perform robust watermark extraction for efficient results with less computational time. Simulation results verify that the scheme has significant imperceptibility, efficient robust watermark extraction, correct authentication and completely reversible nature for ROI recovery. The relative investigation with existing schemes confirms the dominance of the proposed work over already existing work.

Towards perceptual image watermarking with robust texture measurement

Improving the perceptual quality of images is challenging for the robust watermarking method. To address this problem, most existing algorithms proposed to guide the watermark embedding based on just noticeable difference (JND) model, which mainly consists of corresponding contrast masking factor. However, existing metrics are mostly based on the intra-block texture energy, which cannot exactly describe the texture information and meet the needs of watermarking robustness. In this paper, we present a new texture metric, which incorporates the visual saliency based inter-block texture regularity with the robust discrete cosine transform (DCT) coefficients. The JND model is constructed as a modulation factor by using the proposed robust texture metric, which embeds the intra-block energy and inter-block regularity measurement to capture local and global features. Based on the proposed model, we introduce an adaptive quantization step for the watermarking framework. Both quantitative and qualitative results show that our scheme achieves better performance in comparison with several state-of-the-art schemes.

Waveform Design for Watermark Framework Based DFRC System With Application on Joint SAR Imaging and Communication

In this article, the watermarking framework for electromagnetic systems is established with nonblind, semiblind, and blind watermarking demodulation processes mitigated to applications like radar detection, synchronization, integrated dual function, etc. Desired for similar advantages and trade-offs of watermarking technology, the dual function radar and communication (DFRC) system is specifically concerned for covert communication and low possibility of interception (LPI) radar sensing. In this respect, we propose a novel DFRC waveform design method where peak sidelobe level (PSL) of autocorrelation function (ACF) is considered as the figure of merit with information embedded via discrete Fourier transform (DFT) watermarking strategy. Meanwhile, the peak-to-average ratio (PAR) and energy constraints are forced to ensure compatibility with current hardware technique. To handle the resulting NP-hard design problems, the proximal method of multipliers (PMM) is employed with the overall computational burden linear with the amount of information per pulse and quadratic with respect to the code length. Finally, numerical and experimental results are provided to evaluate the effectiveness of the proposed DFRC waveform design scheme with application in joint synthetic aperture radar (SAR) imaging and communication.

Encoded Feature Enhancement in Watermarking Network for Distortion in Real Scenes

Deep-learning based watermarking framework has be extensively studied recently. The main structure of such framework is an encoder, a noise layer and a decoder. By training with different distortion sets in the noise layer, the whole network can realize different robustness. However, such framework has a huge drawback that is the noise layer must be differentiable, otherwise it cannot be trained end-to-end. But for practical use, much distortions are non-differentiable, so such framework cannot be applied. To address such limitations, this paper propose a triple-phase watermarking framework for practical distortions. The proposed framework consists of three phases including a noise-free initial phase, a mask-guided frequency enhancement phase and an adversarial-training phase. Phase 1 aims to initialize an encoder to embed watermark with high visual quality and a decoder to extract the watermark. In order to generate high quality watermarked image, we design the just noticeable difference (JND)-mask image loss in phase 1 to guide the encoder. At phase 2, based on the investigation of the encoded features and distortions, we propose a mask-guided frequency enhancement algorithm to enhance the encoded feature which ensures the survival of such features after distortion, so that there will be enough features to be learned in phase 3. And phase 3 aims to train a stronger decoder to extract the watermark from the image after practical distortions. The combination of these 3 phases can well handle the non-differentiable problems and make the whole network trainable. Various experiments indicate the superior performance of the proposed scheme in the view of traditional differentiable image processing distortion robustness and practical non-differentiable distortion robustness.

FFT-Based Dual-Mode Blind Watermarking for Hiding Binary Logos and Color Images in Audio

Watermarking technology is a feasible approach to prevent copyright infringement. The practicality of the watermark itself affects the acceptance and penetration of relevant applications. In this study, a dual-mode FFT-based audio watermarking framework is devised to achieve both binary and numeric data hiding concurrently using respective modulations. We demonstrate: (1) the adaptive magnitude modulation (AMM) applied to large-magnitude FFT coefficients is competent to deliver efficient and effective binary watermarking at a rate of 678.46 bits per second; (2) the phase modulation (PM) in charge of the embedding of pixel values directly into medium-size FFT coefficients can operate at a rate of 508.85 numbers per second. With a tactical selection of embedding positions and strengths, both the FFT-based AMM and PM exhibit exceptional robustness against commonly encountered attacks. According to the perceptual evaluation of audio quality (PEAQ) metric, the resulting audio quality remains indistinguishable from the original audio, even at the aforementioned high payload capacity.

Image Watermarking Framework Using Histogram Equalization and Visual Saliency

Image Watermarking Framework Using Histogram Equalization and Visual Saliency

A Dual Model Watermarking Framework for Copyright Protection in Image Processing Networks

Image processing networks have gained great success in many fields, and thus the issue of copyright protection for image processing networks has become a focus of attention. Model watermarking techniques are widely used in model copyright protection, but there are two challenges: (1) designing universal trigger sample watermarking for different network models is still a challenge; (2) existing methods of copyright protection based on trigger s watermarking are difficult to resist forgery attacks. In this work, we propose a dual model watermarking framework for copyright protection in image processing networks. The trigger sample watermark is embedded in the training process of the model, which can effectively verify the model copyright. And we design a common method for generating trigger sample watermarks based on generative adversarial networks, adaptively generating trigger sample watermarks according to different models. The spatial watermark is embedded into the model output. When an attacker steals model copyright using a forged trigger sample watermark, which can be correctly extracted to distinguish between the piratical and the protected model. The experiments show that the proposed framework has good performance in different image segmentation networks of UNET, UNET++, and FCN (fully convolutional network), and effectively resists forgery attacks.

A Robust image watermarking for copyright protection using Multiresolution Walsh-Hadamard transform and Schur using Differential evolution

\textit{Background} The proposed watermarking framework is based on the multi-resolution Walsh-Hadamard Transform (MR-WHT) along with Schur factorization. For embedding, the cover image is converted into a frequency domain using MR-WHT before Schur factorization is applied. \textit{Methods} The upper triangle's highest stable eigenvalues in the Schur factorization are employed as reliable embedding locations for the watermark. The singular values of these coefficients are updated with the singular values of the watermark, and the matrix array generated by the biggest eigenvalues is further factorised by singular value decomposition using multiple scaling factors and Differential evolution (DE). Later, an inverse SVD, inverse Schur, and inverse multiresolution Walsh-Hadamard transform is executed to obtain a watermarked image. \textit{Results} The simulation results show that the proposed method is imperceptible and robust against existing watermarking algorithms used in the literature.

The spatial frequency domain designated watermarking framework uses linear blind source separation for intelligent visual signal processing.

This paper develops a digital watermarking algorithm using an informed watermark retrieval architecture. The developed method uses the fractional Fourier transform to embed the watermark in the space-frequency domain and extracts the watermark using blind source separation techniques. The watermark embedding is further enhanced using a heuristic algorithm to increase the strength of the watermarking system. We use genetic algorithm to find the optimal fractional domain by minimizing the coefficient of RMSE between the input image and the watermarked image. The algorithm's performance against various common attacks, e.g., JPEG compression and Gaussian noise, is presented to estimate the algorithm's robustness.

Robust watermarking based on blur‐guided JND model for macrophotography images

Macrophotography Images (MPIs) have recently emerged as an active topic due to the development of mobile phone camera technology. A large number of MPIs have been rapidly increasing in many rich visual services, such as smartphones or high-definition monitors. MPIs are often composed of sharp macroimage and blur background, which exhibit different perceptual properties that often lead to different just noticeable difference (JND) estimation. Inspired by this, we formulate the blur concealment (BC) as another factor to determine the total masking effect: the interaction is relatively straightforward with a limited masking effect in the sharp regions, and is complicated with a strong masking effect in the blur parts. Furthermore, texture and orientation adaption and color information weighting are separately incorporated into the contrast masking and color masking. Finally, considering both BC and masking effects, a novel robust watermarking framework based on the proposed blur-guided JND model for MPIs, targeting at further improving the MPIs copyright protection performance. Extensive experiments on MP2020 and Blur Detection data sets show that the applicability of the proposed JND model in the scenario of perceptually MPIs watermarking, and our proposed scheme can outperform the state-of-the-art watermarking schemes by providing better robustness performance at the uniform visual quality.

QuatJND: A Robust Quaternion JND Model for Color Image Watermarking

Robust quantization watermarking with perceptual JND model has made a great success for image copyright protection. Generally, either restores each color channel separately or processes the vector representation from three color channels with the traditional monochromatic model. And it cannot make full use of the high correlation among RGB channels. In this paper, we proposed a robust quaternion JND Model for color image watermarking (QuatJND). In contrast to the existing perceptual JND models, the advantage of QuatJND is that it can integrate quaternion representation domain and colorfulness simultaneously, and QuatJND incorporates the pattern guided contrast masking effect in quaternion domain. On the other hand, in order to efficiently utilize the color information, we further develop a robust quantization watermarking framework using the color properties of the quaternion DCT coefficients in QuatJND. And the quantization steps of each quaternion DCT block in the scheme are optimal. Experimental results show that our method has a good performance in term of robustness with better visual quality.

Evaluation of a Framework for Robust Image Reversible Watermarking

In the literature, robust reversible watermarking schemes (RWSs) allow the extraction of watermarks after the images have suffered attacks; however, the modified images are compromised. On the other hand, self-recovery schemes will restore the compromised regions of the images, but no secret messages are inserted in these schemes. A framework for robust reversible watermarking with signal restoration capabilities was previously proposed in the literature. This study selects four fragile reversible watermarking techniques and two self-recovery schemes to design different framework configurations. These configurations are evaluated to test the framework’s performance and determine the structure that yields better results in terms of perceptual transparency using a well-known image database as the signal input. It was found that fragile reversible watermarking schemes hold low perceptual distortion, while self-recovery schemes produce high perceptual distortion levels. The inherent characteristics of each algorithm determine, a priori, the behavior of the framework, which is approximated by a proposed equation.

Embedding Guided End-to-End Framework for Robust Image Watermarking

In recent years, deep learning-based watermarking algorithms have received extensive attention. However, the existing algorithms mainly use the autoencoder to insert watermark automatically and ignore using the prior knowledge to guide the watermark embedding. In this paper, an end-to-end framework based on embedding guidance is proposed for robust image watermarking. It contains four modules, i.e., prior knowledge extractor, encoder, attacking simulator, and decoder. To guide the watermark embedding, the prior knowledge extractor providing chrominance and edge information of cover images is used to modify cover images before inserting the watermark by the encoder. To enhance the robustness of watermark extraction, the attacking simulator applying various differentiable attacks on the encoded images is introduced before extracting the watermark by the decoder. Experimental results show that the proposed algorithm achieves a good balance between invisibility and robustness and is superior to state-of-the-art algorithms.

A Visible Video Data Hiding Scheme Based on Fade-In and Fade-Out Effects Utilizing Barcodes

Data hiding techniques are usually applied into digital watermarking or digital fingerprinting, which is used to protect intellectual property rights or to avoid illegal copies of the original works. It has been pointed out that data hiding can be utilized as a communication medium. In conventional digital watermarking frameworks, it is required that the difference between the cover objects and the stego objects are quite small, such that the difference cannot be recognized by human sensory systems. On the other hand, the authors have proposed a ‘hearable’ data hiding technique for audio signals that can carry secret messages and can be naturally recognized as a musical piece by human ears. In this study, we extend the idea of the hearable data hiding into video signals by utilizing the visual effects. As visual effects, we employ fade-in and fade-out effects which can be used as a kind of visual rendering for scene transitions. In the proposed schemes, secret messages are generated as one-dimensional barcodes which are used for fade-in or fade-out effects. The present paper shows that the proposed schemes have the high accuracy in extracting the embedded messages even from the video signals captured by smartphones or tablets. It is also shown that the video signals conveying the embedded messages can be naturally recognized by human visual systems through subjective evaluation experiments.

Spatial Domain-Based Robust Watermarking Framework for Cultural Images

Heritage multimedia, which include photographs, customs, knowledge, arts, rituals, audio, cultural information, and music, are valuable artifacts of any region. The most important attribute of heritage media is the transmission of important features of past generations, which reflect their way of living, innovative attitude, and diversity in archaeological and historical perspectives. However, the proliferation of the Internet has made such data exchange more challenging than ever, allowing unauthorized users to easily access such information. Under such circumstances, securing cultural heritage (CH) media is essential. In that regard, herein, we present a spatial domain-based blind and robust watermarking scheme for the ownership verification of colored CH images; this scheme uses DC coefficient modification. In this scheme, the “Y” element of the YCbCr space is used for inserting a watermark. The “Y” element of a host image is divided into non-overlapping blocks with sizes of 8 × 8. Each 8 × 8 block is then divided into two 4 × 8 subblocks. Instead of calculating the DC coefficients using the discrete cosine transform, we independently calculate the DC coefficient of every 4 × 8 subblock in the spatial domain. We test our method based on standard test images obtained from the USC-SIPI dataset and a self-created dataset of cultural images. Our scheme demonstrates improved robustness and lower computational complexity than frequency-domain-based techniques. The average peak signal-to-noise ratio of the proposed technique for test images is 40.0830 dB, and the structural similarity index matrix value is closer to one under no attack, ensuring the imperceptibility of the technique. Further, we prove the resilience of the proposed algorithm by comparing it with various state-of-the-art techniques.

Re‐shooting Resistant Blind Watermarking Framework Based on Feature Separation With Gaussian Mixture Model

Watermarking is a technology to embed digital data into a digital image. In particular, blind watermarking is expected to ensure the authenticity and/or accountability of digital images without compromising the visual impression of viewing. A re‐shooting resistant blind watermarking method is proposed in this study. This method uses the features of a given image to select regions, which are modified to embed digital data. As the proposed embedding process preserves these image features, no markers are required to extract the embedded data from the watermarked image. Furthermore, based on the estimation of the watermarked image region, the method crops the re‐shot image before the extraction. Therefore, the proposed method works, even if the re‐shot image includes the outer area surrounding the watermarked image region. Moreover, the proposed method compensates for the geometric transformation of the re‐shot image based on the definition of the coordinate system, which is based on the distribution of the image feature points. Therefore, the proposed method works when the re‐shot image is taken diagonally in front of the watermarked image. Experimental results validated that the proposed method can extract the embedded digital data from re‐shot images of a watermarked image under considerable conditions of re‐shooting. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

An optimized framework for digital watermarking based on multi-parameterized 2D-FrFT using PSO

In this paper, we propose a multi-parameter image watermarking algorithm using the fractional Fourier transform (FrFT). The proposed algorithm employs a double transform mechanism to increase the robustness of the algorithm against attacks. We also couple the application of the transform with the Particle Swarm Optimization (PSO) algorithm to decide on the ideal fractional domain for the embedding process. The PSO finds the optimal fractional domain by minimizing the RMSE between the input image and the watermarked image. The results obtained are tested against standard parameters, e.g., robustness, imperceptibility, payload capacity, to evaluate the performance of the algorithm.