FTG: Score-based black-box watermarking by fragile trigger generation for deep model integrity verification

Deep neural networks(DNNs)are widely used in real-world applications,thanks to their exceptional performance in image recognition.However,their vulnerability to attacks,such as Trojan and data poison,can compromise the integrity and stability of DNN applications.Therefore,it is crucial to verify the integrity of DNN models to ensure their security.Previous research on model watermarking for integrity detection has encountered the issue of overexposure of model parameters during embedding and extraction of the watermark.To address this problem,we propose a novel score-based black-box DNN fragile watermarking framework called fragile trigger generation(FTG).The FTG framework only requires the prediction probability distribution of the final output of the classifier during the watermarking process.It generates different fragile samples as the trigger,based on the classification prediction probability of the target classifier and a specified prediction probability mask to watermark it.Different prediction probability masks can promote the generation of fragile samples in corresponding distribution types.The whole watermarking process does not affect the performance of the target classifier.When verifying the watermarking information,the FTG only needs to compare the prediction results of the model on the samples with the previous label.As a result,the required model parameter information is reduced,and the FTG only needs a few samples to detect slight modifications in the model.Experimental results demonstrate the effectiveness of our proposed method and show its superiority over related work.The FTG framework provides a robust solution for verifying the integrity of DNN models,and its effectiveness in detecting slight modifications makes it a valuable tool for ensuring the security and stability of DNN applications.

An Intelligent Sensor Data Preprocessing Method for OCT Fundus Image Watermarking Using an RCNN

Watermarks can provide reliable and secure copyright protection for optical coherence tomography(OCT)fundus images.The effective image segmentation is helpful for promoting OCT image watermarking.However,OCT images have a large amount of low-quality data,which seriously affects the performance of segmentationmethods.Therefore,this paper proposes an effective segmentation method for OCT fundus image watermarking using a rough convolutional neural network(RCNN).First,the rough-set-based feature discretization module is designed to preprocess the input data.Second,a dual attention mechanism for feature channels and spatial regions in the CNN is added to enable the model to adaptively select important information for fusion.Finally,the refinement module for enhancing the extraction power of multi-scale information is added to improve the edge accuracy in segmentation.RCNN is compared with CE-Net and MultiResUNet on 83 gold standard 3D retinal OCT data samples.The average dice similarly coefficient(DSC)obtained by RCNN is 6%higher than that of CE-Net.The average 95 percent Hausdorff distance(95HD)and average symmetric surface distance(ASD)obtained by RCNN are 32.4%and 33.3%lower than those of MultiResUNet,respectively.We also evaluate the effect of feature discretization,as well as analyze the initial learning rate of RCNN and conduct ablation experiments with the four different models.The experimental results indicate that our method can improve the segmentation accuracy of OCT fundus images,providing strong support for its application in medical image watermarking.

Novel self-embedding holographic watermarking image encryption protection scheme

For digital image transmission security and information copyright,a new holographic image self-embedding watermarking encryption scheme is proposed.Firstly,the plaintext is converted to the RGB three-color channel,the corresponding phase hologram is obtained by holographic technology and the watermark is self-embedded in the frequency domain.Secondly,by applying the Hilbert transform principle and genetic center law,a complete set of image encryption algorithms is constructed to realize the encryption of image information.Finally,simulation results and security analysis indicate that the scheme can effectively encrypt and decrypt image information and realize the copyright protection of information.The introduced scheme can provide some support for relevant theoretical research,and has practical significance.

Fractional-order heterogeneous memristive Rulkov neuronal network and its medical image watermarking application

This article proposes a novel fractional heterogeneous neural network by coupling a Rulkov neuron with a Hopfield neural network(FRHNN),utilizing memristors for emulating neural synapses.The study firstly demonstrates the coexistence of multiple firing patterns through phase diagrams,Lyapunov exponents(LEs),and bifurcation diagrams.Secondly,the parameter related firing behaviors are described through two-parameter bifurcation diagrams.Subsequently,local attraction basins reveal multi-stability phenomena related to initial values.Moreover,the proposed model is implemented on a microcomputer-based ARM platform,and the experimental results correspond to the numerical simulations.Finally,the article explores the application of digital watermarking for medical images,illustrating its features of excellent imperceptibility,extensive key space,and robustness against attacks including noise and cropping.

MarkINeRV: A Robust Watermarking Scheme for Neural Representation for Videos Based on Invertible Neural Networks

Recent research advances in implicit neural representation have shown that a wide range of video data distributions are achieved by sharing model weights for Neural Representation for Videos(NeRV).While explicit methods exist for accurately embedding ownership or copyright information in video data,the nascent NeRV framework has yet to address this issue comprehensively.In response,this paper introduces MarkINeRV,a scheme designed to embed watermarking information into video frames using an invertible neural network watermarking approach to protect the copyright of NeRV,which models the embedding and extraction of watermarks as a pair of inverse processes of a reversible network and employs the same network to achieve embedding and extraction of watermarks.It is just that the information flow is in the opposite direction.Additionally,a video frame quality enhancement module is incorporated to mitigate watermarking information losses in the rendering process and the possibility ofmalicious attacks during transmission,ensuring the accurate extraction of watermarking information through the invertible network’s inverse process.This paper evaluates the accuracy,robustness,and invisibility of MarkINeRV through multiple video datasets.The results demonstrate its efficacy in extracting watermarking information for copyright protection of NeRV.MarkINeRV represents a pioneering investigation into copyright issues surrounding NeRV.

MarkNeRF:Watermarking for Neural Radiance Field

This paper presents a novel watermarking scheme designed to address the copyright protection challenges encountered with Neural radiation field(NeRF)models.We employ an embedding network to integrate the watermark into the images within the training set.Then,theNeRFmodel is utilized for 3Dmodeling.For copyright verification,a secret image is generated by inputting a confidential viewpoint into NeRF.On this basis,design an extraction network to extract embedded watermark images fromconfidential viewpoints.In the event of suspicion regarding the unauthorized usage of NeRF in a black-box scenario,the verifier can extract the watermark from the confidential viewpoint to authenticate the model’s copyright.The experimental results demonstrate not only the production of visually appealing watermarks but also robust resistance against various types of noise attacks,thereby substantiating the effectiveness of our approach in safeguarding NeRF.

RWNeRF:Robust Watermarking Scheme for Neural Radiance Fields Based on Invertible Neural Networks

As neural radiance fields continue to advance in 3D content representation,the copyright issues surrounding 3D models oriented towards implicit representation become increasingly pressing.In response to this challenge,this paper treats the embedding and extraction of neural radiance field watermarks as inverse problems of image transformations and proposes a scheme for protecting neural radiance field copyrights using invertible neural network watermarking.Leveraging 2D image watermarking technology for 3D scene protection,the scheme embeds watermarks within the training images of neural radiance fields through the forward process in invertible neural networks and extracts them from images rendered by neural radiance fields through the reverse process,thereby ensuring copyright protection for both the neural radiance fields and associated 3D scenes.However,challenges such as information loss during rendering processes and deliberate tampering necessitate the design of an image quality enhancement module to increase the scheme’s robustness.This module restores distorted images through neural network processing before watermark extraction.Additionally,embedding watermarks in each training image enables watermark information extraction from multiple viewpoints.Our proposed watermarking method achieves a PSNR(Peak Signal-to-Noise Ratio)value exceeding 37 dB for images containing watermarks and 22 dB for recovered watermarked images,as evaluated on the Lego,Hotdog,and Chair datasets,respectively.These results demonstrate the efficacy of our scheme in enhancing copyright protection.

Improving Video Watermarking through Galois Field GF(2^(4)) Multiplication Tables with Diverse Irreducible Polynomials and Adaptive Techniques

Video watermarking plays a crucial role in protecting intellectual property rights and ensuring content authenticity.This study delves into the integration of Galois Field(GF)multiplication tables,especially GF(2^(4)),and their interaction with distinct irreducible polynomials.The primary aim is to enhance watermarking techniques for achieving imperceptibility,robustness,and efficient execution time.The research employs scene selection and adaptive thresholding techniques to streamline the watermarking process.Scene selection is used strategically to embed watermarks in the most vital frames of the video,while adaptive thresholding methods ensure that the watermarking process adheres to imperceptibility criteria,maintaining the video's visual quality.Concurrently,careful consideration is given to execution time,crucial in real-world scenarios,to balance efficiency and efficacy.The Peak Signal-to-Noise Ratio(PSNR)serves as a pivotal metric to gauge the watermark's imperceptibility and video quality.The study explores various irreducible polynomials,navigating the trade-offs between computational efficiency and watermark imperceptibility.In parallel,the study pays careful attention to the execution time,a paramount consideration in real-world scenarios,to strike a balance between efficiency and efficacy.This comprehensive analysis provides valuable insights into the interplay of GF multiplication tables,diverse irreducible polynomials,scene selection,adaptive thresholding,imperceptibility,and execution time.The evaluation of the proposed algorithm's robustness was conducted using PSNR and NC metrics,and it was subjected to assessment under the impact of five distinct attack scenarios.These findings contribute to the development of watermarking strategies that balance imperceptibility,robustness,and processing efficiency,enhancing the field's practicality and effectiveness.

Digital Text Document Watermarking Based Tampering Attack Detection via Internet

Owing to the rapid increase in the interchange of text information through internet networks,the reliability and security of digital content are becoming a major research problem.Tampering detection,Content authentication,and integrity verification of digital content interchanged through the Internet were utilized to solve a major concern in information and communication technologies.The authors’difficulties were tampering detection,authentication,and integrity verification of the digital contents.This study develops an Automated Data Mining based Digital Text Document Watermarking for Tampering Attack Detection(ADMDTW-TAD)via the Internet.The DM concept is exploited in the presented ADMDTW-TAD technique to identify the document’s appropriate characteristics to embed larger watermark information.The presented secure watermarking scheme intends to transmit digital text documents over the Internet securely.Once the watermark is embedded with no damage to the original document,it is then shared with the destination.The watermark extraction process is performed to get the original document securely.The experimental validation of the ADMDTW-TAD technique is carried out under varying levels of attack volumes,and the outcomes were inspected in terms of different measures.The simulation values indicated that the ADMDTW-TAD technique improved performance over other models.

Secure Transmission of Compressed Medical Image Sequences on Communication Networks Using Motion Vector Watermarking

Medical imaging plays a key role within modern hospital management systems for diagnostic purposes.Compression methodologies are extensively employed to mitigate storage demands and enhance transmission speed,all while upholding image quality.Moreover,an increasing number of hospitals are embracing cloud computing for patient data storage,necessitating meticulous scrutiny of server security and privacy protocols.Nevertheless,considering the widespread availability of multimedia tools,the preservation of digital data integrity surpasses the significance of compression alone.In response to this concern,we propose a secure storage and transmission solution for compressed medical image sequences,such as ultrasound images,utilizing a motion vector watermarking scheme.The watermark is generated employing an error-correcting code known as Bose-Chaudhuri-Hocquenghem(BCH)and is subsequently embedded into the compressed sequence via block-based motion vectors.In the process of watermark embedding,motion vectors are selected based on their magnitude and phase angle.When embedding watermarks,no specific spatial area,such as a region of interest(ROI),is used in the images.The embedding of watermark bits is dependent on motion vectors.Although reversible watermarking allows the restoration of the original image sequences,we use the irreversible watermarking method.The reason for this is that the use of reversible watermarks may impede the claims of ownership and legal rights.The restoration of original data or images may call into question ownership or other legal claims.The peak signal-to-noise ratio(PSNR)and structural similarity index(SSIM)serve as metrics for evaluating the watermarked image quality.Across all images,the PSNR value exceeds 46 dB,and the SSIM value exceeds 0.92.Experimental results substantiate the efficacy of the proposed technique in preserving data integrity.

Robust blind image watermarking based on interest points

Digital watermarking technology plays an essential role in the work of anti-counterfeiting and traceability.However,image watermarking algorithms are weak against hybrid attacks,especially geometric attacks,such as cropping attacks,rotation attacks,etc.We propose a robust blind image watermarking algorithm that combines stable interest points and deep learning networks to improve the robustness of the watermarking algorithm further.First,to extract more sparse and stable interest points,we use the Superpoint algorithm for generation and design two steps to perform the screening procedure.We first keep the points with the highest possibility in a given region to ensure the sparsity of the points and then filter the robust interest points by hybrid attacks to ensure high stability.The message is embedded in sub-blocks centered on stable interest points using a deep learning-based framework.Different kinds of attacks and simulated noise are added to the adversarial training to guarantee the robustness of embedded blocks.We use the ConvNext network for watermark extraction and determine the division threshold based on the decoded values of the unembedded sub-blocks.Through extensive experimental results,we demonstrate that our proposed algorithm can improve the accuracy of the network in extracting information while ensuring high invisibility between the embedded image and the original cover image.Comparison with previous SOTA work reveals that our algorithm can achieve better visual and numerical results on hybrid and geometric attacks.

A Dual Domain Robust Reversible Watermarking Algorithm for Frame Grouping Videos Using Scene Smoothness

The proposed robust reversible watermarking algorithm addresses the compatibility challenges between robustness and reversibility in existing video watermarking techniques by leveraging scene smoothness for frame grouping videos.Grounded in the H.264 video coding standard,the algorithm first employs traditional robust watermark stitching technology to embed watermark information in the low-frequency coefficient domain of the U channel.Subsequently,it utilizes histogram migration techniques in the high-frequency coefficient domain of the U channel to embed auxiliary information,enabling successful watermark extraction and lossless recovery of the original video content.Experimental results demonstrate the algorithm’s strong imperceptibility,with each embedded frame in the experimental videos achieving a mean peak signal-to-noise ratio of 49.3830 dB and a mean structural similarity of 0.9996.Compared with the three comparison algorithms,the performance of the two experimental indexes is improved by 7.59%and 0.4%on average.At the same time,the proposed algorithm has strong robustness to both offline and online attacks:In the face of offline attacks,the average normalized correlation coefficient between the extracted watermark and the original watermark is 0.9989,and the average bit error rate is 0.0089.In the face of online attacks,the normalized correlation coefficient between the extracted watermark and the original watermark is 0.8840,and the mean bit error rate is 0.2269.Compared with the three comparison algorithms,the performance of the two experimental indexes is improved by 1.27%and 18.16%on average,highlighting the algorithm’s robustness.Furthermore,the algorithm exhibits low computational complexity,with the mean encoding and the mean decoding time differentials during experimental video processing being 3.934 and 2.273 s,respectively,underscoring its practical utility.

Reversible Semi-Fragile Watermarking Technique for Integrity Control of Relational Database

Reversible watermarking schemes for relational database are usually classified into two groups: robust schemes and fragile schemes. The main limitation of existing reversible fragile methods is that they cannot differentiate between legal and malicious modifications. In this paper, we introduce a novel lossless semi-fragile scheme based on prediction-error expansion for content protection of relational database. In the proposed method, all attributes in a database relation are first classified according to their sensitivity to legitimate updates. Then, the watermark is embedded by expanding the prediction error of the two least significant digits of securely selected attributes. At watermark extraction, the proposed method has the ability to fully restore the original data while detecting and localizing tampering. The applicability of our method is demonstrated theoretically and experimentally.

Boosting image watermarking authenticity spreading secrecy from counting-based secret-sharing

This study presents enhancing images authentication by securing watermarking hidden data via shares generated from counting-based secret sharing.The trustfulness of shares utilised secret-sharing as an applicable privacy creation tool for the authentication of real-life complex platforms.This research adjusts embedding the watermarking data over the images by innovative redistribution of shares to be embedded spread over all the images.The anticipated watermarking technique guaranteed to scatter the share bits implanting at different least significant bits of image pixels as boosting up the trust overall authentication practicality.The paper experimentation performance analysis shows that this improved image watermarking authentication(capacity)is averagely better by 33%–67%than other related exclusive-OR oriented and octagon approaches.Interestingly,these measurement improvements did not degrade the robustness and security of the system,inspiring our research for opening novel track of related future counting-based secret-sharing authentication progresses to come.

New Fragile Watermarking Technique to Identify Inserted Video Objects Using H.264 and Color Features

The transmission of video content over a network raises various issues relating to copyright authenticity,ethics,legality,and privacy.The protection of copyrighted video content is a significant issue in the video industry,and it is essential to find effective solutions to prevent tampering and modification of digital video content during its transmission through digital media.However,there are stillmany unresolved challenges.This paper aims to address those challenges by proposing a new technique for detectingmoving objects in digital videos,which can help prove the credibility of video content by detecting any fake objects inserted by hackers.The proposed technique involves using two methods,the H.264 and the extraction color features methods,to embed and extract watermarks in video frames.The study tested the performance of the system against various attacks and found it to be robust.The evaluation was done using different metrics such as Peak-Signal-to-Noise Ratio(PSNR),Mean Squared Error(MSE),Structural Similarity Index Measure(SSIM),Bit Correction Ratio(BCR),and Normalized Correlation.The accuracy of identifying moving objects was high,ranging from 96.3%to 98.7%.The system was also able to embed a fragile watermark with a success rate of over 93.65%and had an average capacity of hiding of 78.67.The reconstructed video frames had high quality with a PSNR of at least 65.45 dB and SSIMof over 0.97,making them imperceptible to the human eye.The system also had an acceptable average time difference(T=1.227/s)compared with other state-of-the-art methods.

Robust Watermarking Algorithm for Medical Volume Data Based on Polar Cosine Transform and 3D-DCT

The amount of 3D data stored and transmitted in the Internet of Medical Things(IoMT)is increasing,making protecting these medical data increasingly prominent.However,there are relatively few researches on 3D data watermarking.Moreover,due to the particularity of medical data,strict data quality should be considered while protecting data security.To solve the problem,in the field of medical volume data,we proposed a robust watermarking algorithm based on Polar Cosine Transform and 3D-Discrete Cosine Transform(PCT and 3D-DCT).Each slice of the volume data was transformed by PCT to obtain feature row vector,and then the reshaped three-dimensional feature matrix was transformed by 3D-DCT.Based on the contour information of the volume data and the detail information of the inner slice,the visual feature vector was obtained by applying the per-ceptual hash.In addition,the watermark was encrypted by a multi-sensitive initial value Sine and Piecewise linear chaotic Mapping(SPM)system,and embedded as a zero watermark.The key was stored in a third party.Under the same experimental conditions,when the volume data is rotated by 80 degrees,cut 25%along the Z axis,and the JPEG compression quality is 1%,the Normalized Correlation Coefficient(NC)of the extracted watermark is 0.80,0.89,and 1.00 respectively,which are significantly higher than the comparison algorithm.

A Multi-Watermarking Algorithm for Medical Images Using Inception V3 and DCT

Medical images are a critical component of the diagnostic process for clinicians.Although the quality of medical photographs is essential to the accuracy of a physician’s diagnosis,they must be encrypted due to the characteristics of digital storage and information leakage associated with medical images.Traditional watermark embedding algorithm embeds the watermark information into the medical image,which reduces the quality of the medical image and affects the physicians’judgment of patient diagnosis.In addition,watermarks in this method have weak robustness under high-intensity geometric attacks when the medical image is attacked and the watermarks are destroyed.This paper proposes a novel watermarking algorithm using the convolutional neural networks(CNN)Inception V3 and the discrete cosine transform(DCT)to address above mentioned problems.First,the medical image is input into the Inception V3 network,which has been structured by adjusting parameters,such as the size of the convolution kernels and the typical architecture of the convolution modules.Second,the coefficients extracted from the fully connected layer of the network are transformed by DCT to obtain the feature vector of the medical image.At last,the watermarks are encrypted using the logistic map system and hash function,and the keys are stored by a third party.The encrypted watermarks and the original image features are performed logical operations to realize the embedding of zero-watermark.In the experimental section,multiple watermarking schemes using three different types of watermarks were implemented to verify the effectiveness of the three proposed algorithms.Our NC values for all the images are more than 90%accurate which shows the robustness of the algorithm.Extensive experimental results demonstrate the robustness under both conventional and high-intensity geometric attacks of the proposed algorithm.

Adaptive Reversible Visible Watermarking Based on Total Variation for BTC-Compressed Images

Few previous Reversible Visible Watermarking(RVW)schemes have both good transparency and watermark visibility.An adaptive RVW scheme that integrates Total Variation and visual perception in Block Truncation Coding(BTC)compressed domain,called TVB-RVW is proposed in this paper.A new mean image estimation method for BTC-compressed images is first developed with the help of Total Variation.Then,a visual perception factor computation model is devised by fusing texture and luminance characteristics.An adaptive watermark embedding strategy is used to embed the visible watermark with the effect of the visual perception factor in the BTC domain.Moreover,a lossless embedding method of the encrypted visible watermark is exploited to deter illegal watermark removal.The visible watermark can be removed since the visual perception factor and the estimated mean image remain unchanged before and after watermark embedding.Extensive experiments validate the superiority of the proposed algorithm over previous RVW schemes in BTC in terms of the visual quality of watermarked images and watermark visibility,and it can achieve a good balance between transparency and watermark visibility.

Robust Multi-Watermarking Algorithm for Medical Images Based on GoogLeNet and Henon Map

The field of medical images has been rapidly evolving since the advent of the digital medical information era.However,medical data is susceptible to leaks and hacks during transmission.This paper proposed a robust multi-watermarking algorithm for medical images based on GoogLeNet transfer learning to protect the privacy of patient data during transmission and storage,as well as to increase the resistance to geometric attacks and the capacity of embedded watermarks of watermarking algorithms.First,a pre-trained GoogLeNet network is used in this paper,based on which the parameters of several previous layers of the network are fixed and the network is fine-tuned for the constructed medical dataset,so that the pre-trained network can further learn the deep convolutional features in the medical dataset,and then the trained network is used to extract the stable feature vectors of medical images.Then,a two-dimensional Henon chaos encryption technique,which is more sensitive to initial values,is used to encrypt multiple different types of watermarked private information.Finally,the feature vector of the image is logically operated with the encrypted multiple watermark information,and the obtained key is stored in a third party,thus achieving zero watermark embedding and blind extraction.The experimental results confirmthe robustness of the algorithm from the perspective ofmultiple types of watermarks,while also demonstrating the successful embedding ofmultiple watermarks for medical images,and show that the algorithm is more resistant to geometric attacks than some conventional watermarking algorithms.

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 hasbecome a focus of attention. Model watermarking techniques are widely usedin model copyright protection, but there are two challenges: (1) designinguniversal trigger sample watermarking for different network models is stilla challenge;(2) existing methods of copyright protection based on trigger swatermarking are difficult to resist forgery attacks. In this work, we propose adual model watermarking framework for copyright protection in image processingnetworks. The trigger sample watermark is embedded in the trainingprocess of the model, which can effectively verify the model copyright. And wedesign a common method for generating trigger sample watermarks based ongenerative adversarial networks, adaptively generating trigger sample watermarksaccording to different models. The spatial watermark is embedded intothe model output. When an attacker steals model copyright using a forgedtrigger sample watermark, which can be correctly extracted to distinguishbetween the piratical and the protected model. The experiments show that theproposed framework has good performance in different image segmentationnetworks of UNET, UNET++, and FCN (fully convolutional network), andeffectively resists forgery attacks.