Le tatouage d’images (ou watermarking) consiste à insérer de l’information de copyright, ou d’autres messages de vérification, invisible à l’oeil humain. Parmi toutes les approches possibles, le tatouage psychovisuel est le plus performant. Il consiste à utiliser les connaissances du Système Visuel Humain (SVH) pour en déduire la force du tatouage de manière optimale. Dans cette étude, nous avons proposé une nouvelle méthode de tatouage d'image psychovisuel qui combine les avantages des ondelettes à celle du domaine spatial par l’utilisation d'une représentation en ondelettes à échelle mixée (JNDEM). En considérant la densité des coefficients d'ondelettes dominants, notre système est capable de différencier les zones uniformes, celles incluant des contours, et enfin celles texturées. Nous avons sélectionné les méthodes efficaces de tatouage psychovisuel adaptée à chaque zone, à savoir celle de Chou pour les zones uniformes, de nouveau celle de Chou pour les contours, et finalement celle de Qi pour les textures. Nous avons aussi pris en compte la sensibilité du SVH qui varie avec l’orientation locale de l’image. Nous avons comparé notre JNDEM à quatre autres méthodes de tatouage de la littérature en termes de qualité visuelle et de robustesse. La méthode JNDEM conduit à une grande qualité visuelle mesurée par des tests objectifs et subjectifs. De plus elle est robuste aux attaques. Ce dernier point est d'une grande importance pour les applications réelles. En conclusion, la méthode proposée est le meilleur compromis entre la qualité visuelle et la résistance aux attaques comparée aux quatre autres méthodes testées. / Watermarking consists in inserting copyright information, or other verification messages, invisible to the human eye. Among all the possible approaches, perceptual watermarking is the most efficient. Perceptual image watermarking consists in using knowledge of the Human Visual System (HVS) to choose the strength of the watermark. In this study, we propose a new perceptual image watermarking method (JNDEM ) that combines the advantages of both the wavelet domain and the spatial domain since a mixed-scale wavelet representation is applied. By considering the density of the dominant wavelet coefficients, our scheme is able to differentiate uniform, edge and texture areas. We selected effective Just Noticeable Difference (JND) models from the literature, i.e. luminance and edge masking developed by Chou, and texture masking developed by Qi. We also took into account the HVS sensitivity which varies with the orientation of the image activity. We have compared our JNDEM to four other JND models from the literature in terms of visual quality and of robustness to attacks. Results show that the proposed method yields a high visual quality and is very robust against attacks. This last point is of great importance for real applications. In conclusion, the proposed method is the best trade-off between visual quality and resistance to attacks among the four other tested methods.

De nos jours la protection des données numériques est un problème très important. Que ce soit pour des applications de confidentialité, de communication, de traçabilité ou d'identification par exemple, il est nécessaire de développer des techniques adaptées. Dans le cadre de cette thèse en collaboration avec la société STRATEGIES S.A., la méthode choisie pour la protection de maillages 3D est l'insertion de données cachées, également appelée tatouage numérique. Pour des données 3D, un des problèmes les plus importants est la phase de synchronisation qui intervient dans les algorithmes d'insertion et d'extraction des données. Cette phase permet de repérer, de sélectionner et d'ordonner les « zones » qui sont privilégiées pour la dissimulation d'information. Nous avons choisi d'orienter le manuscrit sur cette phase. Ainsi, nous proposons une classification des méthodes de tatouages en fonction de leur méthode de synchronisation. Puis en se basant sur des techniques de synchronisation par des structures de données, telle que les arbres couvrants de poids minimum, nous proposons une analyse théorique de cette structure. Dans un premier temps nous expliquons les raisons de la sensibilité des arbres à la mobilité des points. Puis connaissant ses faiblesses, nous proposons une autre technique de synchronisation toujours basée sur les arbres couvrants de poids minimum. / Data security is one of the main issue in computer science. We need to develop solutions for confidentiality, communication, fingerprinting or identification applications for exemple. In this thesis made with STRATEGIES S.A., the chosen method to protect 3D meshes is watermarking.Watermarking is divided in two steps, the embedding and the extraction. In both of them a synchronization phase is needed. It is one of the most important step for 3D mesh because it permits to look for areas available to embed information, and order them. All the thesis is devoted to the synchronization step. First of all, we propose a classification of watermarking techniques based on the type of synchronization method instead of evaluation criterions such as robustness or capacity.Then, from methods based on Euclidean minimum spanning tree, we propose a theoritical analysis of the mobility of the vertices in that kind of structure. First, we explain the reasons of the sensibility of the structure. Secondly, we propose another scheme based on the Euclidean minimum spanning tree knowing its fragility.

The proliferation of portable systems and mobile computing platforms has increased the need for the design of low power consuming integrated circuits. The increase in chip density and clock frequencies due to technology advances has made low power design a critical issue. Low power design is further driven by several other factors such as thermal considerations and environmental concerns. In low-power design for battery driven portable applications, the reduction of peak power, peak power differential, average power and energy are equally important. In this dissertation, we propose a framework for the reduction of these parameters through datapath scheduling at behavioral level. Several ILP based and heuristic based scheduling schemes are developed for datapath synthesis assuming : (i) single supply voltage and single frequency (SVSF), (ii) multiple supply voltages and dynamic frequency clocking (MVDFC), and (iii) multiple supply voltages and multicycling (MVMC). The scheduling schemes attempt to minimize : (i) energy, (ii) energy delay product, (iii) peak power, (iv) simultaneous peak power and average power, (v) simultaneous peak power, average power, peak power differential and energy, and (vi) power fluctuation. A new parameter called "Cycle Power Function" (CPF) is defined which captures the transient power characteristics as the equally weighted sum of normalized mean cycle power and normalized mean cycle differential power. Minimizing this parameter using multiple supply voltages and dynamic frequency clocking results in the reduction of both energy and transient power. The cycle differential power can be modeled as either the absolute deviation from the average power or as the cycle-to-cycle power gradient. The switching activity information is obtained from behavioral simulations. Power fluctuation is modeled as the cycle-to-cycle power gradient and to reduce fluctuation the mean power gradient (MPG) is minimized. The power models take into consideration the effect of switching activity on the power consumption of the functional units. Experimental results for selected high-level synthesis benchmark circuits under different constraints indicate that significant reductions in power, energy and energy delay product can be obtained and that the MVDFC and MVMC schemes yield better power reduction compared to the SVSF scheme. Several application specific VLSI circuits were designed and implemented for digital watermarking of images. Digital watermarking is the process that embeds data called a watermark into a multimedia object such that the watermark can be detected or extracted later to make an assertion about the object. A class of VLSI architectures were proposed for various watermarking algorithms : (i) spatial domain invisible-robust watermarking scheme, (ii) spatial domain invisible-fragile watermarking scheme, (iii) spatial domain visible watermarking scheme, (iv) DCT domain invisible-robust watermarking scheme, and (v) DCT domain visible watermarking scheme. Prototype implementation of (i), (ii) and (iii) are given. The hardware modules can be incorporated in a "JPEG encoder" or in a "digital still camera".

University of Macau / Faculty of Science and Technology / Department of Computer and Information Science

The multimedia watermarking techniques are evolved very quickly in the last years with the increase in the use of internet. The evolution of the internet makes the copyright issues very important and many different approaches are appeared to protect the digital content. On the other hand, holography is the method to store the 3-D information of an object but it is very applicable to use as a watermark because of the nature of the holographic data. The 3-D object can be reconstructed from the hologram even if the hologram is cropped or occluded. However, watermarking of an image with a hologram is a very novel approach and there are only a few works in the literature which are not very robust against the attacks like filtering or compression. In this thesis, we propose to embed the phase of the hologram to an image using quantization index modulation (QIM). QIM is utilized to make the watermarking scheme blind and degrade the host image as low as possible. The robustness of the proposed technique is also tested against several attacks such as filtering, compression, etc. The evaluated performance of this system is compared with the existing methods in the literature which uses either holograms or logos as the secret mark. Furthermore, the characteristics of the holograms are investigated and the findings about the hologram compression are reported in this work.

The first explored area in this thesis is the proposed data hiding method, TCQ-IS. The method is based on Trellis Coded Quantization (TCQ), whose initial state selection is arbitrary. TCQ-IS exploits this fact to hide data. It is a practical multi-dimensional that eliminates the prohibitive task of designing high dimensional quantizers. The strength and weaknesses of the method are stated by various experiments. The second contribution is the proposed data hiding method, Forbidden Zone Data Hiding (FZDH), which relies on the concept of &ldquo / forbidden zone&rdquo / , where host signal is not altered. The main motive of FZDH is to introduce distortion as much as needed, while keeping a range of host signal intact depending on the desired level of robustness. FZDH is compared against Quantization Index Modulation (QIM) as well as DC-QIM and ST-QIM. FZDH outperforms QIM even in 1-D and DC-QIM in higher dimensions. Furthermore, FZDH is comparable with ST-QIM for certain operation regimes. The final contribution is the video data hiding framework that includes FZDH, selective embedding and Repeat Accumulate (RA) codes. De-synchronization due to selective embedding is handled with RA codes. By means of simple rules applied to the embedded frame markers, certain level of robustness against temporal attacks is introduced. Selected coefficients are used to embed message bits by employing multi-dimensional FZDH. The framework is tested with typical broadcast material against common video processing attacks. The results indicate that the framework can be utilized in real life applications.

Includes vita. / Title from PDF of title page. / Document formatted into pages; contains 289 pages. / Thesis (Ph.D.)--University of South Florida, 2003. / Includes bibliographical references. / Text (Electronic thesis) in PDF format. / ABSTRACT: The proliferation of portable systems and mobile computing platforms has increased the need for the design of low power consuming integrated circuits. The increase in chip density and clock frequencies due to technology advances has made low power design a critical issue. Low power design is further driven by several other factors such as thermal considerations and environmental concerns. In low-power design for battery driven portable applications, the reduction of peak power, peak power differential, average power and energy are equally important. In this dissertation, we propose a framework for the reduction of these parameters through datapath scheduling at behavioral level. Several ILP based and heuristic based scheduling schemes are developed for datapath synthesis assuming : (i) single supply voltage and single frequency (SVSF), (ii) multiple supply voltages and dynamic frequency clocking (MVDFC), and (iii) multiple supply voltages and multicycling (MVMC). / ABSTRACT: The scheduling schemes attempt to minimize : (i) energy, (ii) energy delay product, (iii) peak power, (iv) simultaneous peak power and average power, (v) simultaneous peak power, average power, peak power differential and energy, and (vi) power fluctuation. A new parameter called "Cycle Power Function" CPF) is defined which captures the transient power characteristics as the equally weighted sum of normalized mean cycle power and normalized mean cycle differential power. Minimizing this parameter using multiple supply voltages and dynamic frequency clocking results in the reduction of both energy and transient power. The cycle differential power can be modeled as either the absolute deviation from the average power or as the cycle-to-cycle power gradient. The switching activity information is obtained from behavioral simulations. Power fluctuation is modeled as the cycle-to-cycle power gradient and to reduce fluctuation the mean power gradient MPG is minimized. / ABSTRACT: The power models take into consideration the effect of switching activity on the power consumption of the functional units. Experimental results for selected high-level synthesis benchmark circuits under different constraints indicate that significant reductions in power, energy and energy delay product can be obtained and that the MVDFC and MVMC schemes yield better power reduction compared to the SVSF scheme. Several application specific VLSI circuits were designed and implemented for digital watermarking of images. Digital watermarking is the process that embeds data called a watermark into a multimedia object such that the watermark can be detected or extracted later to make an assertion about the object. / ABSTRACT: A class of VLSI architectures were proposed for various watermarking algorithms : (i) spatial domain invisible-robust watermarking scheme, (ii) spatial domain invisible-fragile watermarking scheme, (iii) spatial domain visible watermarking scheme, (iv) DCT domain invisible-robust watermarking scheme, and (v) DCT domain visible watermarking scheme. Prototype implementation of (i), (ii) and (iii) are given. The hardware modules can be incorporated in a "JPEG encoder" or in a "digital still camera". / System requirements: World Wide Web browser and PDF reader. / Mode of access: World Wide Web.

The thesis presents research on multiobjective optimization based on memetic computing and its applications in engineering. We have introduced a framework for adaptive multiobjective memetic optimization algorithms (AMMOA) with an information theoretic criterion for guiding the selection, clustering, and local refinements. A robust stopping criterion for AMMOA has also been introduced to solve non-linear and large-scale optimization problems. The framework has been implemented for different benchmark test problems with remarkable results. This thesis also presents two applications of these algorithms. First, an optimal image data hiding technique has been formulated as a multiobjective optimization problem with conflicting objectives. In particular, trade-off factors in designing an optimal image data hiding are investigated to maximize the quality of watermarked images and the robustness of watermark. With the fixed size of a logo watermark, there is a conflict between these two objectives, thus a multiobjective optimization problem is introduced. We propose to use a hybrid between general regression neural networks (GRNN) and the adaptive multiobjective memetic optimization algorithm (AMMOA) to solve this challenging problem. This novel image data hiding approach has been implemented for many different test natural images with remarkable robustness and transparency of the embedded logo watermark. We also introduce a perceptual measure based on the relative Rényi information spectrum to evaluate the quality of watermarked images. The second application is the problem of joint spectrum sensing and power control optimization for a multichannel, multiple-user cognitive radio network. We investigated trade-off factors in designing efficient spectrum sensing techniques to maximize the throughput and minimize the interference. To maximize the throughput of secondary users and minimize the interference to primary users, we propose a joint determination of the sensing and transmission parameters of the secondary users, such as sensing times, decision threshold vectors, and power allocation vectors. There is a conflict between these two objectives, thus a multiobjective optimization problem is used again in the form of AMMOA. This algorithm learns to find optimal spectrum sensing times, decision threshold vectors, and power allocation vectors to maximize the averaged opportunistic throughput and minimize the averaged interference to the cognitive radio network. / February 2016

L’avénement des réseaux sociaux, tel que Facebook, MySpace et LinkedIn, a fourni une plateforme permettant aux individus de rester facilement connectés avec leurs amis, leurs familles ou encore leurs collègues tout en les encourageant activement à partager leurs données personnelles à travers le réseau. Avec la richesse des activités disponibles sur un réseau social, la quantité et la variété des informations personnelles partagées sont considérables. De plus, de part leur nature numérique, ces informations peuvent être facilement copiées, modifiées ou divulguées sans le consentement explicite de leur propriétaire. Ainsi, l’information personnelle révélée par les réseaux sociaux peut affecter de manière concrète la vie de leurs utilisateurs avec des risques pour leur vie privée allant d’un simple embarras à la ruine complète de leur réputation, en passant par l’usurpation d’identité. Malheureusement, la plupart des utilisateurs ne sont pas conscients de ces risques et les outils mis en place par les réseaux sociaux actuels ne sont pas suffisants pour protéger efficacement la vie privée de leurs utilisateurs. En outre, même si un utilisateur peut contrôler l’accès à son propre profil, il ne peut pas contrôler ce que les autres révèlent à son sujet. En effet, les “amis” d’un utilisateur sur un réseau social peuvent parfois révéler plus d’information à son propos que celui-ci ne le souhaiterait. Le respect de la vie privée est un droit fondamental pour chaque individu. Nous pré- sentons dans cette thèse une approche qui vise à accroître la prise de conscience des utilisateurs des risques par rapport à leur vie privée et à maintenir la souveraineté sur leurs données lorsqu’ils utilisent un réseau social. La première contribution de cette thèse réside dans la classification des risques multiples ainsi que les atteintes à la vie privée des utilisateurs d’un réseau social. Nous introduisons ensuite un cadre formel pour le respect de la vie privée dans les réseaux sociaux ainsi que le concept de politique de vie privée (UPP). Celle-ci définie par l’utilisateur offre une manière simple et flexible de spécifier et communiquer leur attentes en terme de respect de la vie privée à d’autres utilisateurs, tiers parties ainsi qu’au fournisseur du réseau social. Par ailleurs, nous dé- finissons une taxonomie (possiblement non-exhaustive) des critères qu’un réseau social peut intégrer dans sa conception pour améliorer le respect de la vie privée. En introduisant le concept de réseau social respectueux de la vie privée (PSNS), nous proposons Privacy Watch, un réseau social respectueux de la vie privée qui combine les concepts de provenance et d’imputabilité afin d’aider les utilisateurs à maintenir la souveraineté sur leurs données personnelles. Finalement, nous décrivons et comparons les différentes propositions de réseaux sociaux respectueux de la vie privée qui ont émergé récemment. Nous classifions aussi ces différentes approches au regard des critères de respect de la vie privée introduits dans cette thèse. / The rise of Social Networking Sites (SNS), such as Facebook, Myspace, and LinkedIn has provided a platform for individuals to easily stay in touch with friends, family and colleagues and actively encourage their users to share personal information. With the wealth of activities available on SNS, the amount and variety of personal information shared is considerable and diverse. Additionally, due to its digital nature, this information can be easily copied, modified and disclosed without the explicit consent of their owner. Personal information disclosed from SNS could affect users’ life, with privacy risks ranging from simple embarrassment to ruining their reputation, or even identity theft. Unfortunately, many users are not fully aware of the danger of divulging their personal information and the current privacy solutions are not flexible and thorough enough to protect user data. Furthermore, even though users of SNS can control access to their own profile, they cannot control what others may reveal about them. Friends can sometimes be untrustworthy and disclose more information about the user than they should. Considering that privacy is a fundamental right for every individual, in this thesis, we present an approach that increases privacy awareness of the users and maintains the sovereignty of their data when using SNS. The first contribution of this thesis is the classification of multiple types of risks as well as user expectations regarding privacy in SNS. Afterwards, we introduce the Privacy Framework for SNS and the concept of User Privacy Policy (UPP) to offer users an easy and flexible way to specify and communicate their privacy concerns to other users, third parties and SNS provider. Additionally, we define a taxonomy (possibly non-exhaustive) of privacy criteria that can enhance the user privacy if they are integrated within the design of a SNS and introduce the concept of a Privacy-enhanced SNS (PSNS). Furthermore, we present also Privacy Watch, a theoretical proposal of a PSNS platform that combines the concept of provenance and accountability to help SNS users maintain sovereignty over their personal data. Finally, we survey and compare several privacy-enhanced SNS that were recently proposed that try to integrate some privacy features directly into the design of the system. We also classify these different approaches with respect to the privacy criteria developed.

The present thesis addresses the MPEG-4 AVC stream watermarking and considers two theoretical and applicative challenges, namely ownership protection and content integrity verification.From the theoretical point of view, the thesis main challenge is to develop a unitary watermarking framework (insertion/detection) able to serve the two above mentioned applications in the compressed domain. From the methodological point of view, the challenge is to instantiate this theoretical framework for serving the targeted applications. The thesis first main contribution consists in building the theoretical framework for the multi symbol watermarking based on quantization index modulation (m-QIM). The insertion rule is analytically designed by extending the binary QIM rule. The detection rule is optimized so as to ensure minimal probability of error under additive white Gaussian noise distributed attacks. It is thus demonstrated that the data payload can be increased by a factor of log2m, for prescribed transparency and additive Gaussian noise power. A data payload of 150 bits per minute, i.e. about 20 times larger than the limit imposed by the DCI standard, is obtained. The thesis second main theoretical contribution consists in specifying a preprocessing MPEG-4 AVC shaping operation which can eliminate the intra-frame drift effect. The drift represents the distortion spread in the compressed stream related to the MPEG encoding paradigm. In this respect, the drift distortion propagation problem in MPEG-4 AVC is algebraically expressed and the corresponding equations system is solved under drift-free constraints. The drift-free shaping results in gain in transparency of 2 dB in PSNR