In this paper, a new watermarking system for copyright protection of digital images is presented. The method operates in the frequency domain, by embedding a pseudo-random sequence of real numbers in a selected set of DFT coefficients of the image. Moreover, the masking characteristics of the Human Visual System are exploited for watermark hiding and a synchronization pattern is introduced into the watermarked image to cope with geometrical attacks. By relying on statistical decision theory, a new decoding algorithm which is optimum for non-additive full-frame DFT watermarks is then derived. Experimental results highlight both the superiority of the novel detector scheme with respect to conventional correlation-based decoding, both the robustness of the overall system against a large set of attacks aiming at removing the watermark.

This paper presents a transform domain allocation method for watermarking of digital multimedia content; specific analytical results are obtained for audio watermarking. To solve the problem, the watermarking transmitter, channel and receiver are formally defined and watermarking is described in a detection theory framework. Describing the channel by a generic attack model and using detection theory, the challenge is to minimize the probability of detection error at the receiver. Using strong assumptions on the original and watermark signals to derive analytical results, an optimal allocation method is deduced. The allocation strategy is compared to strategies in most present systems. Moreover, the assumptions on the signals reveal to be realistic in a given implementation of the watermarking system. The allocation method is being patented for digital multimedia content.

This paper presents an other way of considering watermarking methods, which are analyzed from the point of view of the Information Theory. Watermarking is thus a communication problem in which some information bits have to be transmitted through an additive noise channel subjected to distortions and attacks. Designing watermarking methods in such a way that this channel is Gaussian can be profitable. This paper shows to what extent error protection techniques extensively studied for digital communication through Gaussian channels can be used advantageously for watermarking. Convolutional codes combined with soft-decision decoding are the best example. Especially, when soft-decision Viterbi decoding is employed, this kind of coding schemes can achieve much better performance than BCH codes, at comparable levels of complexity and redundancy, both for still and moving images.

Quantization index modulation (QIM) methods, a class of digital watermarking and information embedding methods, achieve very efficient trade-offs among the amount of embedded information (rate), the amount of embedding-induced distortion to the host signal, and the robustness to intentional and unintentional attacks. For example, we show that against independent additive Gaussian attacks, which are good models for at least some types of uniformed and unintentional attacks, QIM methods exist that achieve the best possible rate-distortion-robustness trade-offs (i.e., capacity) asymptotically at high rates and achieve performance within a few dB of capacity at all finite rates. Furthermore, low- complexity realizations of QIM methods, such as so-called dither modulation, have also been shown to achieve favorable rate-distortion-robustness trade-offs. We further develop preprocessing and postprocessing techniques that enable QIM to fully achieve capacity, not only against Gaussian attacks but also against other types of attacks as well. One practical postprocessing technique we develop we refer to as distortion compensation. Distortion compensation has the property that when suitably optimized it is sufficient for use in conjunction with QIM to achieve capacity against Gaussian attacks and against square-error distortion-constrained attacks. More generally, we present the results of a comparative information theoretic analysis of the fundamental performance limits of QIM, distortion-compensated QIM, and other watermarking methods and demonstrate practically achievable gains with experimental results.

A watermarking scheme for distinguishing different copies of the same multimedia document (fingerprinting) is investigated. Efficient transmission of natural data requires lossy compression, which might impair the embedded watermark. We investigate whether the quantization step in compression schemes can be replaced by dithered quantization to combine fingerprinting and compression. Dithered quantization offers the possibility of producing perceptually equivalent signals that are not exactly equal. The non-subtractive quantization error can be used as the watermark. We denote the proposed watermarking scheme as 'quantization watermarking.' Such a scheme is only practical for watermarking applications where the original signal is available to the detector. We analyze the influence of the dither signal on the perceptual quality of the watermarked document and the watermark detection robustness. Further, the cross-talk between the non- subtractive quantization errors for two different dither realizations is investigated. An analytical description for fine quantization and experimental results for coarse quantization show how the cross-talk depends on the characteristics of the dither signal. The derived properties of quantization watermarking are verified for combined JPEG compression and fingerprinting. The detection robustness for the proposed quantization error watermark is compared with that of an independent additive watermark.

One method of attacking an imbedded invisible watermark is to create a derivative image that is geometrically distorted relative to the original. One attack, developed at Cambridge University, is called 'StirMark.' Image-distorting methods modify images so subtly that the changes are essentially unnoticeable to a viewer. However, their effect on invisible watermarks can be devastating, rendering them unextractable. In this paper, an automated countermeasure to image-distorting attacks will be described. Employing an unmarked copy of the original image as a reference, the possible distortion in a suspect image is first detected by the method, then measured, and finally reversed, producing a restored image approximately geometrically aligned with the original. Using a robust invisible watermarking method presented previously by one of the authors to produce a watermarked image, 'StirMark' to distort the watermarked image, and a copy of the original unmarked image for reference, the restoration method is demonstrated to be sufficient by showing successful extraction of the imbedded watermark from a restored image.

In this paper we report a novel method to estimate the scaling factor of a previously scaled watermarked image and the angle by which the image has been rotated. Scaling and rotation performed on a watermarked image, as part of the attacks the image may undergo, can very easily confuse the decoder unless it rescales and/or rotates the image back to its original size/orientation, i.e., recover the lost synchronism. To be able do so, the decoder needs to know by how much the image has been scaled and rotated, i.e., needs to know both the scaling factor and the rotation angle. In our approach, we compute the Edges Standard Deviation Ratio (ESDR) which gives us an accurate estimate for the scaling factor. The rotation angle is approximated by the Average Edges Angles Difference (AEAD). Both ESDR an AEAD are computed from wavelet maxima locations which have been estimated form the non orthogonal dyadic wavelet transform. The proposed scheme does not require the original image provided that a proper normalization has been attained. Our method has proved its robustness to wide rotation and scale ranges.

Many electronic watermarks for still images and video content are sensitive to geometric distortions. For example, simple rotation, scaling, and/or translation (RST) of an image can prevent detection of a public watermark. In this paper, we propose a watermarking algorithm that is robust to RST distortions. The watermark is embedded into a 1-dimensional signal obtained by first taking the Fourier transform of the image, resampling the Fourier magnitudes into log-polar coordinates, and then summing a function of those magnitudes along the log-radius axis. If the image is rotated, the resulting signal is cyclically shifted. If it is scaled, the signal is multiplied by some value. And if the image is translated, the signal is unaffected. We can therefore compensate for rotation with a simple search, and for scaling by using the correlation coefficient for the detection metric. False positive results on a database of 10,000 images are reported. Robustness results on a database of 2,000 images are described. It is shown that the watermark is robust to rotation, scale and translation. In addition, the algorithm shows resistance to cropping.

Numerical information is volatile and watermarking is a solution to assist copyright protection. During the detection step, the synchronization of the mark is a great problem. Geometric transformations can defeat the detector of the mark by desynchronizing the mark. Our scheme is based on warping of pre-defined triangular patterns. The content of the image (feature points) is used to mark independently different regions. This allows the synchronization of the mark for the detection step. Feature points mixed with a Delaunay tessellation permits to mark each triangle of the image. The detection is performed by warping triangle to a reference pattern and correlating with a reference triangle. Different algorithms have been developed in the spatial domain and in the frequential (DCT) domain. Our results show that our schemes are robust to Stirmark and other geometric transformations on different categories of images.

The technology of a digital watermarking that can embed copyright information in valuable images has come to be paid attention. Therefore, we propose a new digital watermarking method using the Smearing-Desmearing transformation and show the algorithm. The Smearing-Desmearing transformation is a new transformation that watermark information is spread on the whole of an image. In our method, we can easily calculate how much the method degrades the quality of the watermarked image from embedded watermark energy.

The recent development of digital multimedia communications together with the intrinsic capability of digital information to be copied and manipulated requires new copyright protection and content authentication schemes to be developed. This paper is devoted to the second issue, the one of image or video content authentication. A computationally efficient spatial watermarking technique for authentication of visual information, robust to small distortions caused by compression, is described. In essence, a content-dependent authentication data is embedded into the picture, by modifying the relationship of image projections throughout the entire image. To obtain a secure data embedding and extraction procedure, directions onto which image parts are projected depend on a secret key. In order to guarantee minimum visibility of the embedded data, the insertion process is used in conjunction with perceptual models, exploiting spatial domain masking effects. The viability of the method as a means for protecting the content is assessed under JPEG compression and semantic content modifications. With the present system, robustness to JPEG compression up to compression factors of about 1:10, can be achieved, maintaining the subjective image quality after watermark insertion. At the same time, it is possible to detect and localize small image manipulations.

In this paper, we present two tamper-detection techniques. The first is a fragile technique that can detect the most minor changes in a marked image using a DCT-based data hiding method to embed a tamper-detection mark. The second is a semi-fragile technique that detects the locations of significant manipulations while disregarding the less important effects of image compression and additive channel noise. Both techniques are fully described and the performance of each algorithms demonstrated by manipulation of the marked images.

In this paper, we propose a semi-fragile watermarking technique that accepts JPEG lossy compression on the watermarked image to a pre-determined quality factor, and rejects malicious attacks. The authenticator can identify the positions of corrupted blocks, and recover them with approximation of the original ones. In addition to JPEG compression, adjustments of the brightness of the image within reasonable ranges, are also acceptable using the proposed authenticator. The security of the proposed method is achieved by using the secret block mapping function which controls the signature generating/embedding processes. Our authenticator is based on two invariant properties of DCT coefficients before and after JPEG compressions. They are deterministic so that no probabilistic decision is needed in the system. The first property shows that if we modify a DCT coefficient to an integral multiple of a quantization step, which is larger than the steps used in later JPEG compressions, then this coefficient can be exactly reconstructed after later acceptable JPEG compression. The second one is the invariant relationships between two coefficients in a block pair before and after JPEG compression. Therefore, we can use the second property to generate authentication signature, and use the first property to embed it as watermarks. There is no perceptible degradation between the watermarked image and the original. In additional to authentication signatures, we can also embed the recovery bits for recovering approximate pixel values in corrupted areas. Our authenticator utilizes the compressed bitstream, and thus avoids rounding errors in reconstructing DCT coefficients. Experimental results showed the effectiveness of this system. The system also guaranies no false alarms, i.e., no acceptable JPEG compression is rejected.

Digital watermarking research has generally focused upon two classes of watermarks, fragile and robust. Robust watermarks are designed to be detected even after attempts are made to remove them. Fragile watermarks are used for authentication purposes and are capable of detecting even minute changes of the watermarked content. Unfortunately, neither type of watermark is ideal when considering 'information preserving' transformations (such as compression) which preserve the meaning or expression of the content and 'information altering' transformations (such as feature replacement) which change the expression of the content. In this paper we describe a semi-fragile watermark for still images that can detect information altering transformations even after the watermarked content is subjected to information preserving alterations.

Authentication techniques provide a means of ensuring the integrity of a message. The recent proliferation of multimedia content has led to a need for developing authentication mechanisms. Although, authentication techniques have been studied for many decades, multimedia authentication poses some new challenges. Perhaps the key such challenge being the need to authenticate multimedia content as opposed to its representation. In this paper, we review some of the techniques proposed in the literature for multimedia content authentication. We then propose distortion bounded authentication techniques that give hard guarantees on the amount of distortion that will be tolerated before the multimedia object under consideration is deemed unauthentic. The basic idea behind distortion-bounded authentication is simple. Quantization is performed (in feature space) before authentication, thereby restricting image features in a known and deterministic manner. The same quantization is performed prior to verification. Distortions less than half the quantization step size will not affect the verification process and the content will be deemed authentic. The basic framework is simple and can be applied with many different techniques, distortion measures and feature sets. We give examples of distortion-bound authentication techniques using the L1 and L2 norms in pixel domain.

A variety of robust non-perceptual watermarking methods have been proposed to solve the illegal copying and proof of ownership problems. Although a number of broad claims have been made in the field of robustness of various digital watermarking methods. It is still difficult to handle combined or non-linear geometric transformations. In this paper we propose the SSP approach. We use a correlation based watermark pattern approach in the spatial domain which operates on image inherent properties when selecting the marking positions: self spanning patterns (SSP). Each watermark pattern which represents one information bit of the whole watermark string is spanned over four image inherent properties like image edges or corners. These four points specify a polygon over which the watermark pattern is spanned using a polygon filling algorithm. After transformations we search for our points based on the image characteristics. Then we span our pattern over the points and check if there was a one or a zero embedded. Continuing that for all pattern positions we can build the whole watermark string. The advantage of the SSPs is that transformations need not be recognized, the marking positions and our pattern shape can be automatically found.

In this paper, we describe some of the problems associated with watermarking key management, with particular attention to the case of video. We also describe a possible solution to the problem, which is that of image-dependent watermarking, and briefly discuss some of the possible advantages to be gained from adopting such an approach. The paper also presents a simple, efficient means of robustly extracting bits from a video sequence. The algorithm has applications to secure, oblivious video watermark detection.

This work discusses implementation issues of real-time video/image/signal processing applications on personal computers. We give a list of performance optimization guidelines and demonstrate some examples by optimizing our video watermark detection scheme. In many applications, watermarking technology must have (1) the ability to be implemented at low cost, (2) robustness against common image processing operations, and (3) resilience against purely malicious attacks. Many works, including ours, have demonstrated watermark robustness and invisibility. This work demonstrates that, after some performance optimizations, we can decode a 704 X 480 MPEG-2 video and detect the watermarks, both in software, and display the decoded video frames in real-time on an Intel Pentium^R III 500 MHz system. Currently, there is only 10.5% overhead of the watermark detection over video decoding. The cost of our optimized implementation is 43% lower than that of the unoptimized version. The optimization techniques demonstrated in this work can be applied to other watermarking schemes and other video/image/signal processing applications.

This paper presents an adaptive video watermarking using motion information. Because video data have one more dimension than image data, simple adaptation of image watermarking method to video data would reveal some types of visual artifact such as flickering. In the proposed scheme, same watermark information is embedded for same region in each frame to decrease such visual artifact. For higher robustness and invisibility, watermarking strength is adjusted by motion information and region complexity.

Invisible watermark images can be incorporated in printed halftone images using specially designed halftone screens. The watermark information is embedded into the image by varying the spatial correlation of the halftone texture. The halftone screen with embedded watermarks can be used exactly as a normal halftone screen, so there is no additional image processing required for processing individual images to embed watermarks. Once the binary output image is printed on the paper, the correlation of the binary image is converted into physical spatial correlation between neighboring areas of the printed image. This correlation relation is not visible to the eye but it can be detected by scanning the printed image on a desktop scanner and processing the scanned image. Printer and scanner distortions can interfere with the self-alignment of the scanned image, so localized adjustments are made to detect the embedded spatial correlation information in the watermarked image. The retrieval of this watermark is robust to copying and distortion and it can be detected in reproductions of the halftone image.

With the ease of distribution of digital images, there is a growing concern for copyright control and authentication. While there are many existing watermarking and data hiding methods for natural images, almost none can be applied to halftone images. In this paper, we proposed two novel data hiding methods for halftone images. The proposed Data Hiding Pair-Toggling (DHPT) hides data by forced complementary toggling at pseudo-random locations within a halftone image. It is found to be very effective for halftone images with relatively coarse textures. For halftone images with fine textures (such as error diffusion with Steinberg kernel), the proposed Data Hiding Error Diffusion (DHED) gives significantly better visual quality by integrating the data hiding into the error diffusion operation. Both DHPT and DHED are computationally very simple and yet effective in hiding a relatively large amount of data. Both algorithms yield halftone images with good visual quality.

JPEG is a common image format in the WWW and can potentially be used to hide data for secure internet communication and watermark for copyright control. In this paper, we propose an algorithm to embed the secret or watermark information. The proposed algorithm, named Watermarking by DC Coefficients Modification (WDCM), assumes that the quality factor used in JPEG compression is known. We observe that it is perceptually undetectable if the DC coefficients in certain texture-rich blocks are modified by a small amount. We thus embed the secret information as a binary bit sequence in the quantized DC coefficients in those texture rich blocks. The watermark embedding process can be applied in compression domain without re-encode the data. The information bits are randomized by some pseudo-random noise (PN) sequences, the keys of which are needed for the decoding of the secret information. By embedding the information in the DC components, the proposed algorithm is robust to common JPEG compression if the quality factor is known.

A digital home network is a cluster of digital audio/visual (A/V) devices including set-top boxes, TVs, VCRs, DVD players, and general-purpose computing devices such as personal computers. The network may receive copyrighted digital multimedia content from a number of sources. This content may be broadcast via satellite or terrestrial systems, transmitted by cable operators, or made available as prepackaged media (e.g., a digital tape or a digital video disc). Before releasing their content for distribution, the content owners may require protection by specifying access conditions. Once the content is delivered to the consumer, it moves across home the network until it reaches its destination where it is stored or displayed. A copy protection system is needed to prevent unauthorized access to bit streams in transmission from one A/V device to another or while it is in storage on magnetic or optical media. Recently, two fundamental groups of technologies, encryption and watermarking, have been identified for protecting copyrighted digital multimedia content. This paper is an overview of the work done for protecting content owners' investment in intellectual property.

This paper introduces the concept of Smart Images and explains the use of watermarking technology in their implementation. A Smart Image is a digital or physical image that contains a digital watermark, which leads to further information about the image content via the Internet, communicates ownership rights and the procedure for obtaining usage rights, facilitates commerce, or instructs and controls other computer software or hardware. Thus, Smart Images, empowered by digital watermarking technology, act as active agents or catalysts which gracefully bridge both traditional and modern electronic commerce. This paper presents the use of Digimarc Corporation's watermarking technology to implement Smart Images. The paper presents an application that demonstrates how Smart Images facilitate both traditional and electronic commerce. The paper also analyzes the technological challenges to be faced for ubiquitous use of Smart Images.

Information technology has made major strides in the past decade. As results, there have been widespread applications of data storage and transmission. The valuable multimedia information in digital forms, however, is vulnerable to unauthorized access while in storage and during the transmission. Network security and image encryption become important and high profile issues. Image encryption requires manipulating massive amounts of data at high speeds. The use of software in image encryption provides flexibility for manipulation but may not meet some timing constraints. In this paper, we present a novel technique for image encryption using block cipher. The primary concept is based on the implementation of 3-Way encryption algorithm. Experiment results show that our proposed method significantly enhances the security for transmission images over network as well as for storage. Beside the software implementation, we present the hardware implementation of 3-way image encryption algorithm based on FPGA technology. With the flexibility of software implementation and the high performance of microprocessor, FPGA-based cryptosystem is a promising technology for the future network security.

Digital watermarks have recently been proposed for authentication and fingerprinting of both video data and still images and for integrity verification of visual multimedia. In such applications, the watermark must be oblivious and has to depend on a secret key and on the original image. It is important that the dependence on the key be sensitive, while the dependence on the image be continuous (robust). Both requirements can be satisfied using special image digest (hash) functions that return the same bit-string for whole class of images derived from an original image using common processing operations including rotation and scaling. It is further required that two completely different images produce completely different bit-strings. In this paper, we extend our previous work on robust image digest functions describing ideas how to make the hash function independent of image orientation and size. The robust image digest can be clearly utilized for other applications, such as a search index for an efficient image database search.

Digital watermarks can be classified into two categories according to the embedding and retrieval domain, i.e. spatial- and frequency-domain watermarks. Because the two watermarks have different characteristics and limitations, combination of them can have various interesting properties when applied to different applications. In this research, we examine two spatial-frequency composite watermarking schemes. In both cases, a frequency-domain watermarking technique is applied as a baseline structure in the system. The embedded frequency- domain watermark is robust against filtering and compression. A spatial-domain watermarking scheme is then built to compensate some deficiency of the frequency-domain scheme. The first composite scheme is to embed a robust watermark in images to convey copyright or author information. The frequency-domain watermark contains owner's identification number while the spatial-domain watermark is embedded for image registration to resist cropping attack. The second composite scheme is to embed fragile watermark for image authentication. The spatial-domain watermark helps in locating the tampered part of the image while the frequency-domain watermark indicates the source of the image and prevents double watermarking attack. Experimental results show that the two watermarks do not interfere with each other and different functionalities can be achieved. Watermarks in both domains are detected without resorting to the original image. Furthermore, the resulting watermarked image can still preserve high fidelity without serious visual degradation.

The fundamental conditions that have to be fulfilled by a cryptographically secure watermark generation algorithm can be summarized as follows: availability of a large number of different watermarking keys, confusion to ensure that properties of any watermark key are not reflected in the corresponding watermark, and diffusion demanding that similar watermark keys always do result in completely different watermarks. Due to several drawbacks associated with methods based on Anosov diffeomorphisms we propose to use discrete versions of chaotic Kolmogorov flows instead to increase the cryptanalytic strength of the watermark generation scheme. For this approach it can be shown by combinatorial arguments that the number of different watermark keys grows exponentially and that fulfillment of the confusion requirement can be checked by standard statistical tests delivering results which are definitely within a reasonable experimental error when compared to the confidence level in use, while simulations investigating the diffusion distance amount to an average diffusion distance that is perfectly in line with the optimum average diffusion distance calculated analytically.

Our previous work introduced a stochastic framework for studying watermarking. When the receiver used direct correlation detection and the attack employed linear shift- invariant (LSI) filtering, the optimal attack consisted of Wiener filtering and adding noise. Resisting the Wiener attack led to a power-spectrum condition (PSC): the watermark and original power spectra should be directly proportional. PSC- compliant watermarks are most difficult to estimate from the watermarked document in the minimum mean-squared error sense. This paper investigates the fundamental limits of PSC- compliant watermarks; it has two main parts. First, expressions are derived for the detection-error probabilities after the Wiener attack with PSC-compliant Gaussian noise. An explicit relationship between attacked-document distortion and capacity results. The second part studies optimal LSI- filtering attacks and receivers. Because direct correlation detection is only optimal for white Gaussian noise (WGN), the Wiener attack and PSC may actually be suboptimal. It is shown that the PSC holds for a suboptimal, effective white-noise attack. The optimum attack is derived, but there does not appear to be an analytical solution for the optimum watermark power spectrum. Numerical experiments suggest that, as a rule of thumb, white watermarks perform well at lower distortions, while PSC-compliant watermarks perform better at higher distortions.

Watermarking of digital property has received a significant amount of attention in recent years. Robust watermarking of digital images and video for the purpose of copyright protection is a particularly important and challenging topic. In this paper, we provide a technique to improve the robustness of algorithms that require selection of transform coefficients for watermarking. Experimental results show that by strategically choosing the best locations for watermark embedding, the overall robustness of the watermarking algorithm can be improved.

After a brief reminder on the real difficulties that digital watermarking software still has to tackle -- especially some random geometric attacks such as StirMark -- we present an early overview of on-going solutions to make the survival of the watermark possible.

These last years, the rapidly growing digital multimedia market has revealed an urgent need for effective copyright protection mechanisms. Therefore, digital audio, image and video watermarking has recently become a very active area of research, as a solution to this problem. Many important issues have been pointed out, one of them being the robustness to non-intentional and intentional attacks. This paper studies some attacks and proposes countermeasures applied to videos. General attacks are lossy copying/transcoding such as MPEG compression and digital/analog (D/A) conversion, changes of frame-rate, changes of display format, and geometrical distortions. More specific attacks are sequence edition, and statistical attacks such as averaging or collusion. Averaging attack consists of averaging locally consecutive frames to cancel the watermark. This attack works well for schemes which embed random independent marks into frames. In the collusion attack the watermark is estimated from single frames (based on image denoising), and averaged over different scenes for better accuracy. The estimated watermark is then subtracted from each frame. Collusion requires that the same mark is embedded into all frames. The proposed countermeasures first ensures robustness to general attacks by spread spectrum encoding in the frequency domain and by the use of an additional template. Secondly, a Bayesian criterion, evaluating the probability of a correctly decoded watermark, is used for rejection of outliers, and to implement an algorithm against statistical attacks. The idea is to embed randomly chosen marks among a finite set of marks, into subsequences of videos which are long enough to resist averaging attacks, but short enough to avoid collusion attacks. The Bayesian criterion is needed to select the correct mark at the decoding step. Finally, the paper presents experimental results showing the robustness of the proposed method.

Digital image watermarking has become a popular technique for authentication and copyright protection. For verifying the security and robustness of watermarking algorithms, specific attacks have to be applied to test them. In contrast to the known Stirmark attack, which degrades the quality of the image while destroying the watermark, this paper presents a new approach which is based on the estimation of a watermark and the exploitation of the properties of Human Visual System (HVS). The new attack satisfies two important requirements. First, image quality after the attack as perceived by the HVS is not worse than the quality of the stego image. Secondly, the attack uses all available prior information about the watermark and cover image statistics to perform the best watermark removal or damage. The proposed attack is based on a stochastic formulation of the watermark removal problem, considering the embedded watermark as additive noise with some probability distribution. The attack scheme consists of two main stages: (1) watermark estimation and partial removal by a filtering based on a Maximum a Posteriori (MAP) approach; (2) watermark alteration and hiding through addition of noise to the filtered image, taking into account the statistics of the embedded watermark and exploiting HVS characteristics. Experiments on a number of real world and computer generated images show the high efficiency of the proposed attack against known academic and commercial methods: the watermark is completely destroyed in all tested images without altering the image quality. The approach can be used against watermark embedding schemes that operate either in coordinate domain, or transform domains like Fourier, DCT or wavelet.

Research in digital watermarking has progressed along two paths. While new watermarking technologies are being developed, some researchers are also investigating different ways of attacking digital watermarks. Common attacks to watermarks usually aim to destroy the embedded watermark or to impair its detection. In this paper we propose a conceptually new attack for digitally watermarked images. The proposed attack does not destroy an embedded watermark, but copies it from one image to a different image. Although this new attack does not destroy a watermark or impair its detection, it creates new challenges, especially when watermarks are used for copyright protection and identification. The process of copying the watermark requires neither algorithmic knowledge of the watermarking technology nor the watermarking key. The attack is based on an estimation of the embedded watermark in the spatial domain through a filtering process. The estimate of the watermark is then adapted and inserted into the target image. To illustrate the performance of the proposed attack we applied it to commercial and non-commercial watermarking schemes. The experiments showed that the attack is very effective in copying a watermark from one image to a different image. In addition, we have a closer look at application dependent implications of this new attack.

Digital audio watermarking embeds inaudible information into digital audio data for the purposes of copyright protection, ownership verification, covert communication, and/or auxiliary data carrying. In this paper, we first describe the desirable characteristics of digital audio watermarks. Previous work on audio watermarking, which has primarily focused on the inaudibility of the embedded watermark and its robustness against attacks such as compression and noise, is then reviewed. In this research, special attention is paid to the synchronization attack caused by casual audio editing or malicious random cropping, which is a low-cost yet effective attack to watermarking algorithms developed before. A digital audio watermarking scheme of low complexity is proposed in this research as an effective way to deter users from misusing or illegally distributing audio data. The proposed scheme is based on audio content analysis using the wavelet filterbank while the watermark is embedded in the Fourier transform domain. A blind watermark detection technique is developed to identify the embedded watermark under various types of attacks.

An audio registration method based on dynamic time-warping (DTW) technique was described in this paper. DTW technique can register two audio signals whose type of misalignment in time domain is unknown. By measuring the frame dissimilarities in two audio signals, the best alignment between a pair of audio signals can be obtained to register the two audio signals. This method can solve the registration problem for audio signal which is processed by re-scaling in the time domain. It is useful in many audio applications and has been applied to digital audio watermarking detection.

The need to authenticate multimedia content (audio, image and video) brings new challenges to the traditional digital signature schema. Unavoidable transmission distortion of multimedia data, such as noises introduced in printing a certificate, demands new authentication techniques that are resistant to reasonable modifications. In this paper, a novel approach to multimedia authentication based on the vector quantization (VQ) technique is presented. Its security is guaranteed by adopting a typical digital signature scheme and a hierarchical authentication procedure. The flexibility of this approach is achieved by the scalability of VQ. In addition, a feature-based registration procedure is used to pre-filter and restore the received media with reference to the original one. Using this method, a continuous measure of the authenticity can be computed based on the proposed registration and authentication procedure. This measure is helpful for setting a proper threshold for some specific authentication applications. Our experimental results with audio, image, and video have shown that the scheme is robust and practical.

In previous work, Wong had proposed both secret key and public key watermarking schemes for image authentication that can detect and localized any change made to a watermarked image. The techniques proposed were block-based, that is, they partitioned the image into non-overlapping blocks and separately authenticated each block. Subsequently, Holliman and Memon observed that many block based watermarking schemes are vulnerable to substitution attacks. They specifically showed that the Wong schemes can be attacked using a 'vector quantization' (VQ) approach. This attack exploits that fact that if a sufficient number of images containing the same watermark bitmap is available, then one can use a VQ-like technique to forge a watermark into a new image. About the same time and independently, Coppersmith et al. proposed to use overlapping blocks to resist this attack. Although this method can make the attack inefficient, it does so with a significant loss of the localization property of the watermark. We extend in this paper the Wong schemes so that the resulting algorithms can resist the VQ attack and at the same time provide the same localization property in the watermark as the original schemes. The key idea is to insert a unique image-dependent block ID into the watermarking process so that the VQ attack will not have a rich enough 'codebook' to forge the watermark.

In this paper, we describe new and improved attacks on the authentication scheme previously proposed by Yeung and Mintzer. Previous attacks assumed that the binary watermark logo inserted in an image for the purposes of authentication was known. Here we remove that assumption and show how the scheme is still vulnerable, even if the binary logo is not known but the attacker has access to multiple images that have been watermarked with the same secret key and contain the same (but unknown) logo. We present two attacks. The first attack infers the secret watermark insertion function and the binary logo, given multiple images authenticated with the same key and containing the same logo. We show that a very good approximation to the logo and watermark insertion function can be constructed using as few as two images. With color images, one needs many more images, nevertheless the attack is still feasible. The second attack we present, which we call the 'collage-attack' is a variation of the Holliman-Memon counterfeiting attack. The proposed variation does not require knowledge of the watermark logo and produces counterfeits of superior quality by means of a suitable dithering process that we develop.

This paper presents a secure method to handle digital photographs in insurance claim process. It consists of: Device authentication by secure protocol capable to protect photographs from substitution, Watermarking technique to embed digital signatures on JPEG data and detect altered portions, and Robust watermarking technique to embed image identification number which persists throughout decompression/compression operations. The device authentication is used in transmissions of photograph images from a digital camera (DC) to a compact flash memory (CF), and from a CF to a personal computer (PC). Thus images are authenticated at the one-way pass of DC, CF and PC. In registering images on PC, an image identification number and a digital signature are embedded into a JPEG data stream by a watermarking technique called DataHiding^TM. We could imbed 96 bits of data into a 640 X 480 pixel image without degrading visual image quality and persistency of embedding data. After the watermarking process, a photograph can be identified by decoding an embedded number and detected any altered portion by an embedded signature. Using the new method, we have developed a prototype secure claim process, and verified the feasibility of the method at field trials.

With the increasing number of digital transmissions, there is an opportunity to utilize existing broadcast channels to send hidden video transmissions. This presentation outlines and demonstrates a technique to insert a hidden video channel within a larger standard video feed, or 'hide in plain sight.' The approach outlined consists of three parts: data (or video) insertion, data security, and data recovery. These three components are essential to complete a secured digital transmission. To insert the data, knowledge of the standard (for example, MPEG 1, 2, or 4) is used to place the information to be hidden into the compressed data stream post quantization. Since the encoder has perfect knowledge of the receiver's model, it is possible to place the data in the compressed data in a manner that is visually un-intrusive. The technique to hide the data will be presented and an example given. Since, the data (or video) to be inserted may be of a sensitive nature, it is necessary to provide a level of security, which will render the video stream useless if its presence is detected. Both, spatial and temporal dispersion are used to reduce the likelihood of detection with the addition of a modulation of the signal to reduce the image to noise unless the key is known. The intended viewer of the embedded data (or video stream), with knowledge of the key recovers the data by encoding the incoming video stream with the appropriate standard, and extracting the data prior to quantization. While the work in this paper focuses on the insertion of hidden video into a broadcast signal, the same technique serves to prevent tampering to the broadcast signal, since modification of the image data will corrupt the hidden signal. In this way, this provides a level of tamper protection/detection for video, which may be used in legal proceedings.

Audio and video watermarking enable the copyright protection with owner or customer authentication and the detection of media manipulations. The available watermarking technology concentrates on single media like audio or video. But the typical multimedia stream consists of both video and audio data. Our goal is to provide a solution with robust and fragile aspects to guarantee authentication and integrity by using watermarks in combination with content information. We show two solutions for the protection of audio and video data with a combined robust and fragile watermarking approach. The first solution is to insert a time code into the data: We embed a signal as a watermark to detect gaps or changes in the flow of time. The basic idea uses numbers increasing by one. If in the verification process the next number is smaller than the last one or the step is greater than one, the time flow has been changed. This is realized without the combination of video and audio data. But we can synchronize the two data streams. A time signal is only valid if the combination of audio and video signals satisfy a certain attribute. To keep the basic example: if we embed an increasing a number in the audio and a decreasing number in the video, we could test if the combination of both always equals zero. The second solution is more complex: We use watermarks to embed information in each media about the content of the other media. With the help of speech recognition technology it is possible to embed the spoken text, the content, of an audio file in the video. With an algorithm previously developed in [1] we extract video content representation which is embedded in the audio stream. In our paper we present the problem of copyright protection and integrity checks for combined video and audio data. We show our two solutions and discuss our results.

The recent finalization of MPEG-4 will make this standard very attractive for a large range of applications such as video editing, internet video distribution, wireless video communications. Some of these applications will need to address the problem of copyright protection. One of the key points of the MPEG4 standard is the possibility to access and manipulate objects within a video sequence. Thus object watermarking has to be achieved in such a way that, while a video object is transferred from a sequence to another, still it is possible to correctly access the copyright data related to the object itself. The algorithm proposed in this paper embeds a watermark in each video object in a way which is somewhat similar to that presented in reference one. A particular relationship is imposed between some predefined pairs of quantized DCT coefficients in the luminance blocks of pseudo-randomly selected MBs. Watermarks are equally embedded into intra and inter MBs. Experimental results are presented validating the effectiveness of the proposed approach.

This paper investigates two methods for digital watermarking of MPEG2 video sequence. The watermark is spread by Frey's pseudo-chaotic sequence prior to embedding and it could be retrieved by a correlator from the watermarked video without the use of the original video. Both spatial- and DCT (Discrete Cosine Transform)-domain watermark are considered. They are embedded into every I or all I, P and B frames of the MPEG2 video. Some watermark parameters are adjusted to investigate the trade-off between watermark robustness and video quality or capacity of the watermarking system. The quality of watermark is measured by its BER (bit error rate) while the fidelity of watermarked video is measured by its PSNR (Peak Signal to Noise Ratio) as well as by visual inspection. Various forms of cropping attack are found to be detrimental to watermark retrieval but acceptable to the eyes in most cases. In particular, a new form of video watermark attack, i.e. frame cropping, is introduced. Their effects on the watermark robustness are studied and discussed.

Watermarks not specifically designed to be visible should be imperceptible to maintain image or video quality. Watermark characteristics that affect visibility include watermark magnitude, the visual masking model used to determine relative magnitude values, watermark frequency content, and the degree of temporal variation of the watermark between frames. Video characteristics that affect the visibility of a watermark include subject and camera motion, color, resolution, texture, and patterns. We have investigated the effects of these characteristics on watermark visibility for MPEG-1, MPEG-2, and both uncompressed and compressed high-definition video. The average watermark strength that can be added to a video is determined by increasing the average watermark magnitude through the sequence in order to observe the point at which the watermark becomes perceptible. The first artifact to become visible depends on video and watermark characteristics.

Image index tables values generally give the best possible representation of the color information of the image. However, no consideration is given to the arrangement of the color table itself. Thus, depending on the image, pixels with similar colors may have different index values and can therefore have considerably different index binary makeups. Consequently, regions of similarly colored indexed pixels can be noise-like at the bitplane level while the output colors themselves may imply simple bitplane patterns. BPCS image steganography hides information in images based on the principle that if regions in a bitplane are noise-like, those regions can be replaced with noise-like secret data. Therefore, applying traditional BPCS steganography to indexed image data results in drastic visible changes to the image. To overcome this problem, we used a self-organizing neural network to reorder the index table, based on samples from the image, such that similar colors in the index table are near each other with respect to their index values. As a result, regions with similar color information have only slight binary differences at the bitplane level, whereas regions with mixed color information will have considerable binary differences. Using this technique, we can embed secret data that is 15 to 35 percent the size of the image with little or no noticeable degradation in the image.

In this paper we present a novel watermarking scheme to generalize a previous proposal by the same authors. In that paper to watermark, the image amounts to process the colors in the picture as points in the Color Opponency space and to offset each one of them by a random vector. As an extra constraint, in order to avoid picture quality degradation, the offset applied should be such that a color is not moved outside of a small sphere. The watermarking algorithm processes the colors in the picture as points in the Color Opponency space and offsets each one of them, inside a suitable imperceptive area, by a random vector. To improve robustness we suggest to partition the whole pixels population into 'color sets': each set gathers pixels with the same color. Each color set, exceeding a given cardinality, is in turn randomly partitioned into three subsets. Each one of these subsets is hence manipulated in different suitable ways. The watermark is obtained maintaining a record of the statistical distribution of the three subsets. Rigorous theoretical statistical analysis shows that this approach is robust with respect to the most common operations of image processing.

In this paper, a genetic algorithm (GA) based watermark optimization technique for digital image watermarking is proposed. In our approach, watermark embedding positions are simulated by chromosomes, and several GA optimization operators such as reproduction, crossover, and mutation are used. The experimental result shows that the proposed approach can effectively improve the quality of the watermarked image and the robustness of the embedded watermark against various attacks. By analyzing the distribution of embedding positions, we found that watermark embedding positions are highly correlated to the energy distribution of block-DCT coefficients.

Image watermarking has been the topic of much research and it is now widely accepted that the watermark should be placed in perceptually significant components of an image, in order to be robust against common signal distortions and malicious attacks. This choice makes the watermark potentially visible, therefore only slight modifications to the original image are allowed with a low loss of quality. In this work we propose a scheme where the energy of the watermark embedded depends on the image to be marked, thus fitting image capacity. The three fundamental stages of our scheme are: a watermark sequence generation as zero-mean, unit-variance Gaussian noise, an embedding procedure working in the DWT domain, and watermark detection performed by correlation and not requiring the original image. The experimental results prove that the method is resistant to JPEG compression up to a quality factor of 10%, to blurring and median filtering, to cropping and stretching; moreover the watermark is not perceivable for many different kinds of images, including cartoons.

In this paper, we present a method for protection of digital contents by using the watermark embedding in special object, especially, human faces. To insert the watermark signals that are composed of noise like binary signals, we first localize the face regions within images by using the color and edge information. The skin color area is filtered out and then edge detector is applied for skin area to find out face features. These features are used for decision whether the skin area is face region or not. The face region is divide non-overlapping sub-blocks and a watermark bit is inserted into the each sub- block by considering the block activity. We insert a watermark bit in DCT domain of each sub-block. The level of modification of the DCT coefficients is determined considering the block variance. The non-zero coefficients of the DCT are selected and modified according to the robustness levels. Then, inverse DCT is performed. The extraction of the watermark is performed by comparing the original image in DCT domain. The robustness of the watermarking is similar to the other methods in DCT, but it has good visual qualities and less intended external piracy in terms of psychology.