Search

International audience; Synchronization in 3D data hiding is one of the main problems. We need to know where we can embed informa- tion, and be able to find this space in order to extract the message. Various algorithms propose synchronization techniques by triangle or vertex path in a 3D mesh. In this paper, we proposed a new synchronization technique based on Euclidean minimum spanning tree computing (EMST) and the analysis of the displacement of the vertices without moving the connections in the tree. Based on the analysis of the vertices, we select the most robust vertices and synchronize these areas by computing a new EMST called "robust EMST". Then, we analyze the robustness of the technique, i.e. the stability of the most robust vertices selection; and demonstrate the consistence of the criterion selection with the vertex displacement.

International audience; In recent years, a variety of chaos-based cryptosystems have been proposed. Some of these systems are used in designing a pseudo random bit generator (PRBG) for stream cipher applications. Most of the chaotic systems used in cryptography have good chaotic properties like ergodicity, sensitivity to initial values and sensitivity to control parameters. However, some of them are not very suitable for use in cryptography because of their non-uniform density function, and their relatively small key space. To be used in cryptography, a PRBG may need to meet stronger requirements than for other applications. In particular, various statistical tests can be applied to the outputs of such generators to conclude whether the generator produces a truly random sequence or not. In this paper, we propose a PRBG based on the use of the standard chaotic map with large key space and the Engle Continued Fractions (ECF) map. The outputs of the standard map are used as the inputs of ECF-map. The chaotic nature of the standard map and the good statistical properties of the ECF map motivate us to design a new PRBG for stream cipher applications. The numerical simulation analysis indicates that our PRBG produces bit sequences possessing excellent statistical and cryptographic properties.

Secret sharing between two or more parties, exploiting multiplicative homomorphic properties of RSA, may result in false data blocks during the extraction of message. This paper investigates factors that lead to such false blocks and suggest a controlled solution by analyzing their probability distribution. In this paper we prove that false blocks do exist and show that larger sizes of the prime product reduce the error probability of false blocks in extracted messages.

The main objective of this paper is to introduce and illustrate a new tool stemming from Young measure theory in order to capture concentrations, jump sign and gradient oscillations of sequences of SBV-functions. We show how this notion of measure can be applied for the analysis of approximating solutions of Mumford-Shah type energy functionals in the one dimensional case.

International audience; In this paper, we present a novel approach for active finger-printing of state of the art video codec H.264/AVC. Tardos probabilistic fingerprinting code is embedded in H.264/AVC video signals using spread spectrum watermarking technique. Different linear and non-linear collusion attacks have been performed in the pixel domain to show the robustness of the proposed approach. The embedding has been performed in the non-zero quantized transformed coefficients (QTCs) while taking into account the reconstruction loop.

International audience; In this paper, a novel method for fast protection of AVS video coding standard along with compression is presented. Here the problems of compression and selective encryption (SE) have been simultaneously addressed for AVS part-2 Jizhun profile. It is performed in the context-based 2D variable length coding (C2DVLC) module of video codec. SE is performed by using the Advanced Encryption Standard (AES) algorithm with the Cipher Feedback (CFB) mode on a subset of codewords. C2DVLC serves the purpose of encryption step without affecting the coding efficiency of AVS by keeping the bitrate unchanged, generating completely compliant bitstream and utilizing negligible computational power. Nine different benchmark video sequences containing different combinations of motion, texture and objects are used for experimental evaluation of the proposed algorithm.

Perceptual image hashing maps an image to a fixed length binary string based on the image's appearance to the human eye, and has applications in image indexing, authentication, and watermarking. In this paper, we analyze the robustness of perceptual signatures generated by the known SHA1 hash function in the DWT domain. This is achieved by generation signatures from the coefficients of the subband LL after a DWT transformation. Experiments on large-scale database show that the proposed method is efficient, robust against common content-preserving manipulations.

International audience; We propose an adaptively synchronous scalable spread spectrum (A4S) data-hiding strategy to integrate disparate data, needed for a typical 3-D visualization, into a single JPEG2000 for- mat file. JPEG2000 encoding provides a standard format on one hand and the needed multiresolution for scalability on the other. The method has the potential of being imperceptible and robust at the same time. While the spread spectrum (SS) methods are known for the high robustness they offer, our data-hiding strategy is removable at the same time, which ensures highest possible visualization qual- ity. The SS embedding of the discrete wavelet transform (DWT)- domain depth map is carried out in transform domain YCrCb com- ponents from the JPEG2000 coding stream just after the DWT stage. To maintain synchronization, the embedding is carried out while taking into account the correspondence of subbands. Since security is not the immediate concern, we are at liberty with the strength of embedding. This permits us to increase the robustness and bring the reversibility of our method. To estimate the maximum tolerable error in the depth map according to a given viewpoint, a human visual system (HVS)-based psychovisual analysis is also presented.

International audience; In this paper, we propose a new scheme which performs both lossless compression and encryption of data. The lossless compression is based on the arithmetic coding (AC) and the encryption is based on a pseudo random bit generator (PRBG). Thus, the plaintext is compressed with a binary arithmetic coding (BAC) whose two mapping intervals are swapped randomly by using a PRBG. In this paper, we propose a PRBG based on the standard chaotic map and the Engel Continued Fraction (ECF) map to generate a keystream with both good chaotic and statistical properties. To be used in cryptography, a PRBG may need to meet stronger requirements than for other applications. In particular, various statistical tests can be applied to the outputs of such generators to conclude whether the generator produces a truly random sequence or not. The numerical simulation analysis indicates that the proposed compression and encryption scheme satisfies highly security with no loss of the BAC compression efficiency.

In this paper we present a new approach for sharing a secret image between l users exploiting additive homomorphic property of Paillier algorithm. With a traditional approach, when a dealer wants to share an image between l players, the secret image must be sequentially encrypted l + 1 times using l + 1 keys (secret or public keys). When the dealer and the l players want to extract the secret image, they must decrypt sequentially, keeping the same order of the encryption step, by using l + 1 keys (secret or private). With the proposed approach, during the encryption step, each player encrypts his own secret image using the same public key given by the dealer, the dealer encrypts the secret image to be shared with the same key and then the l secret encrypted images plus the encrypted image to be shared are multiplied between them to get a scrambled image. After this step, the dealer can securely use the private key to decrypt this scrambled image to get a new scrambled image which corresponds to the addition of the l + 1 original images because of the additive homomorphic property of Paillier algorithm. When the l players want to extract the secret image, they do not need the dealer and to use keys. Indeed, with our approach, to extract the secret image, the l players need only to subtract their own secret image from the scrambled image. In this paper we illustrate our approach with an example of a captain who wants to share a secret treasure map between l pirates. Experimental results and security analysis show the effectiveness of the proposed scheme.

In 3D steganography and watermarking, the synchronization of the hidden data is a major problem. We need to know where the message is embedded in order to extract the correct information. Various algorithms have been proposed for the last couple of years and we focused on a method based on Euclidean minimum spanning tree (EMST) for the mesh vertices. In this paper we analyze the sensitivity of the EMST structure in order to propose a new method more robust. We present a new theoretical analysis and we propose to visualize the robustness of the EMST. Moreover, we can apply this analysis to various applications that can be useful in 3D steganography such fragile area detection and prediction of the 3D object robustness during transmission on a noisy channel.

In this paper we propose a strategy for seamless tessellation, of varying resolution tiles, based on smoothing and mosaicking in the DWT domain. The scenario involves a tessellation with three different tile qualities or levels of detail (LOD), at a given instant, depending on the viewpoint distance, the time of rendering and hardware resources. The LOD is dependent on the multiresolution characteristic of wavelets from the now widely accepted JPEG2000 codec. Taking the change in viewpoint focus, analogous to a window sliding approach, we believe that at worst the window may come up to be composed of three different tile qualities with the resultant artifacts at tile interfaces. To dilute these artifacts, we treat the tiles at the subband level, in the DWT domain, by employing operations involving suitable subband-sized composite masks, conceived with smoothing and mosaicking in perspective. The resultant composite subbands are subjected to a global inverse DWT to get the final seamless tessellation.

With todays geobrowsers, the tessellations are far from being smooth due to a variety of reasons: the principal being the light difference and resolution heterogeneity. Whilst the former has been extensively dealt with in the literature through classic mosaicking techniques, the latter has got little attention. We focus on this latter aspect and present two DWT domain methods to seamlessly stitch tiles of heterogeneous resolutions. The first method is local in that each of the tiles that constitute the view, is subjected to one of the three context-based smoothing functions proposed for horizontal, vertical, and radial smoothing, depending on its localization in the tessellation. These functions are applied at the DWT subband level and followed by an inverse DWT to give a smoothened tile. In the second method, though we assume the same tessellation scenario, the view field is thought to be of a sliding window which may contain parts of the tiles from the heterogeneous tessellation. The window is refined in the DWT domain through mosaicking and smoothing followed by a global inverse DWT. Rather than the traditional sense, the mosaicking employed over here targets the heterogeneous resolution. Perceptually, this second method has shown better results than the first one. The methods have been successfully applied to practical examples of both the texture and its corresponding DEM for seamless 3D terrain visualization.

International audience; This first decade of twenty-first century is witnessing revolution in the form of memory and network speeds as well as computing efficiencies. Simultaneously, the platform and client diversity base is also expanding, with powerful workstations on one extreme and handheld portable devices, like smart-phones on the other. The range of networks, even if we narrow it down, may end up at optical fiber to EDGE. Still versatile are the clients whose needs evolve with every passing moment. The technological revolution notwithstanding, dealing the accompanying diversity is always a serious challenge that must certainly require some scalable strategies. When it comes to scalability, one is compelled to think of the multiresolution characteristic of wavelets as far as multimedia data is concerned. With the wavelets comes the JPEG2000 standard which has added newer dimensions to the concept of scalability. The scalability challenge is more glaring in the case of applications where disparate and large data is involved. An example application is the area of 3D visualization that requires more than one set of the data. Here, issues like data security, data authentication and data unification then emerge. All these form part and parcel of the field of data hiding. The emergence of the wavelet oriented JPEG2000 codec has brought with it the challenge of when, where and how to hide the data for watermarking or steganography while keeping in view the requirements of hiding capacity, robustness and imperceptibility. Even before the advent of the codec, wavelet transformations were commonplace in the field of signal processing. For the last few years, several techniques to embed data inside a host signal in the wavelet transform domain have been proposed. Few of the methods are compatible with the JPEG2000 standard and that is why one comes across methods exclusively proposed for the codec after the latter's inception. In today's world, data hiding is not limited to just watermarking, steganography or authentication but some nontraditional prospects are also there. One such area is the field of 3D visualization wherein data hiding can be employed as a tool for the synchronous and scalable unification of the disparate data in the JPEG2000 coding pipeline. The rest of this chapter is arranged as follows. Section 2 is dedicated to the background concepts wherein a brief account of the wavelet transformations is being given, followed by the presentation of the salient features of the JPEG2000 codec. After introducing information hiding concepts, Section 3 focuses on the when and where to embed in the JPEG2000 coding pipeline, which leads us to introduce a context level classification of the JPEG2000-based data hiding methods presented in the literature in Section 4. In Section 5 we present an application of the JPEG2000-based data hiding for a scalable and synchronous 3D visualization in a client/server environment. Section 6 sums up this chapter.

International audience; This paper presents a new selective and scalable encryption (SSE) method for intra dyadic scalable coding framework based on wavelet/subband (DWTSB) for H.264/AVC. It has been achieved through the scrambling of quantized transform coefficients (QTCs) in all the subbands of DWTSB. To make the encryption scalable, it takes advantage of the prior knowledge of the frequencies which are dominant in different high frequency (HF) subbands, as traditional zigzag scan is not that efficient for them. Thus, by scrambling the scan order of QTCs in the intra scalable coding framework of H.264/AVC, we were able to get encryption and compression for enhancement layers (ELs) simultaneously. Watermarking has been integrated in the proposed architecture to avoid the requirement of separate keys for each spatial layer. The algorithm is better suited for multimedia streaming as the bitrate of the encrypted bitstream is lower than the original bitrate. Besides offering SSE, the proposed algorithm, when applied to different benchmark video sequences, outperformed the standard zigzag scan in terms of bitrate.

International audience; An adaptive spread spectrum synchronous scalable(A4S) data hiding strategy is being put forward to integrate the disparate 3D facial visualization data, into a single JPEG2000 format file with the aim to cater diverse clients in various bandwidth scenarios. The method is both robust and imperceptible in the sense that the robustness of the spread spectrum (SS) is coupled with the removable embedding that ensures highest possible visualization quality. The SS embedding of the DWT-domain 2.5D facial model is carried out in the transform domain YCrCb components, of the 2D texture, from the coding stream of JPEG2000 codec just after the DWT stage. High depth map quality is ensured through the adaptation of synchronization during embedding that would exclude some highest frequency subbands. The results show that the method can be exploited for video-surveillance and video-conference applications.

International audience; Few work have been proposed for image color protection. A color protection algorithm should give some comprehensive information about the image (such as its grey-level version) but should hide securely the color information. The proposed algorithms are thus specific but take into account recent signal processing concepts such as quantization, compression and data-hiding. Figure 1 illustrates a possible use of the color protection where only secret key owners may rebuild the color image.

International audience; Airborne lidar systems (ALS) provide 3D point clouds of the topography by direct time measurement of a short laser pulse after reflection on the Earth surface. For the last decade, this technique has proved to be the ideal remote sensing tool for delivering very accurate digital terrain model (DTM) of the Earth surface, and then for answering main environmental issues such as natural hazard prevention and natural ressource management. Moreover, such active systems, also called "multiple echo lidar", allow to detect several return signals for a single laser shot. It is particularly relevant in case of vegetation areas since a single lidar survey allows to acquire not only the canopy top (the only visible layer from passive sensors), but also points inside the vegetation layer and on the ground underneath. Thus, among the different remote sensing techniques, airborne laser scanning has also proved to be the most efficient technique to characterize both forest structure and ground topography. For a few years, new airborne laser scanning systems called full-waveform lidar systems have emerged, providing not only 3D point clouds as classical ALS systems, but entire altimeter profiles of reflected energy from the Earth surface. These profiles represent the laser backscattered energy as a function of time. They give to the end-user more control and flexibility on the signal processing steps and enable to extract more information than classical multi-echo lidar data. A detailed state-of-the-art of such systems can be found in [1]. However, managing these data with spacial and time dependency is much more complex than images or 3D point clouds : raw full-waveform lidar data are sets of range profiles of various lengths that are stored in the sensor geometry following both the scan angle of the lidar system and the chronological order along the flight track. Moreover, the data volume is drastically larger than 3D point clouds: it takes about 140 GB for an acquisition time of 1.6 h with a pulse repetition frequency (PRF) of 50kHz. Furthermore, there is neither commercial nor opensource toolkit to handle full-waveform lidar data, but some constructor solutions, that are black boxes, can only extract 3D point clouds from raw data and are designed to their own sensors. Finally, there is not standard file format for full-waveform data (such as the LAS format for multi-echo data). Managing full-waveform lidar data is therefore a challenging task, and we adress this issue by developping a specific research tool: FullAnalyze.

International audience; In this paper, we are particulary interested by image protection. The conventional encryption standard may be not applicable to images, due to the digital images properties which are characterized with some intrinsic features such as highly pixel redundancy and correlation. We propose a new asynchronous stream cipher based on generalized continued fraction (GCF). The proposed scheme is resistant to statistic attack, differential attack and some other known attacks. Experimental results prove that our scheme is efficient and secure.

International audience; A classic approach to obtain High Dynamic Range Images (HDRI) consists in combining multiple images of the same scene with varying exposures. However, if the scene is not static during the time of capture, moving objects will appear blurry and ghosted, i.e. in multiple locations. Detecting and removing ghosting artifacts is an important issue for automatic generation of HDRI of dynamic scenes. This paper first describes a new method for detecting regions where ghosting occur based on an order relation between pixel values in consecutive images. Secondly, we propose to use a robust statistics tool to remove the detected ghosts. Results with different scenes and comparison with two other methods show the effectiveness of the approach.

International audience; This paper presents a novel method for the protection of copyrighted multimedia content. The problem of selective encryption (SE) is being addressed along with the compression for the state of the art video codec H.264/AVC. SE is performed in the context-based adaptive binary arithmetic coding (CABAC) module of video codec. For this purpose, CABAC is converted to an encryption cipher. It has been achieved through scrambling of equal length bin strings. In our scheme, CABAC engine serves the purpose of encryption cipher without affecting the coding efficiency of H.264/AVC by keeping exactly the same bit rate, generating completely compliant bitstream and utilizing negligible computational power. Owing to no escalation in bit rate, our encryption algorithm is better suited for real-time multimedia streaming. It is perfect for playback on handheld devices because of negligible increase in processing power. Nine different benchmark video sequences containing different combinations of motion, texture and objects are used for experimental evaluation of the proposed algorithm.

International audience; This paper presents a novel method for the protection of copyrighted multimedia content. The problem of selective encryption (SE) has been addressed alongwith compression for the state of the art video codec H.264/AVC. SE is performed in the context-based adaptive binary arithmetic coding (CABAC) module of video codec. For this purpose, CABAC is converted to an encryption cipher. It has been achieved through scrambling of equal length binarized code words. In our scheme, CABAC engine serves the purpose of encryption cipher without affecting the coding efficiency of H.264/AVC by keeping exactly the same bitrate and by generating completely compliant bitstream, and requires insignificant computational cost. Nine different benchmark video sequences containing different combinations of motion, texture and objects are used for experimental evaluation of the proposed algorithm.

International audience; The works presented in this paper show how encryption algorithms provide security to medical imagery. The main objective is to guarantee the protection of medical images during transmission, and also once this digital data is archived. The subsequent challenge is to ensure that such coding withstands severe treatment such as compression. For ethical reasons, medical imagery cannot be sent when such a risk is present, and has to be better protected. Encryption is the best form of protection in cases such as this. Many different techniques for the encryption of text already exist. In this paper we show how essential it is to ensure the security of medical imagery and data.

K. HAYAT, W. PUECH and G. SUBSOLLIRMM, UMR CNRS 5506, University of Montpellier II,161, rue Ada, 34392 Montpellier CEDEX 05, FranceE-mail: khizar.hayat@lirmm.fr, william.puech@lirmm.fr and gerard.subsol@lirmm.frwww.lirmm.frG. GESQUIERELSIS, UMR CNRS 6168, Aix-Marseille UniversityIUT, rue R. Follereau, 13200 Arles, FranceE-mail: gilles.gesquiere@lsis.orgwww.lsis.orgWe present a scalable encoding strategy for the 3D facial data in various band-width scenarios. The scalability, needed to cater diverse clients, is achievedthrough the multiresolution characteristic of JPEG2000. The disparate 3D fa-cial data is synchronously uni ed by the application of data hiding wherein the2.5D facial model is embedded in the corresponding 2D texture in the discretewavelet transform (DWT) domain. The uni ed le conforms to the JPEG2000standard and thus no novel format is introduced. The method is e ective andhas the potential to be applied in videosurveillance and videoconference appli-cations.Keywords : JPEG2000, discrete wavelet transform (DWT), scalability, synchro-nization, videosurveillance, videoconference, data hiding, LSB, 3D, 2.5D, facedata, multiresolution

International audience; This paper proposes a new method for the protection of copyrighted multimedia content. Here the problems of compression and selective encryption (SE) have been simultaneously addressed for the state of the art video codec H.264/AVC. SE is performed in the context adaptive variable length coding (CAVLC) module of video codec. For this purpose, CAVLC is converted to an encryption cipher using permutation of equal length codes from a specific variable length coding (VLC) table. In our scheme, entropy coding engine serves the purpose of encryption step without affecting the coding efficiency of H.264/AVC by keeping the bitrate unchanged and generating completely compliant bit stream. Our scheme requires a negligible computational power. Nine different benchmark video sequences containing different combinations of motion, texture and objects are used for experimental evaluation of the proposed algorithm.

The amount of digital visual data (image, video and 3D object) has increased rapidly on the Internet. Image, video and 3D object security becomes increasingly important for many applications, e.g., confidential transmission, video surveillance, military and medical applications. For example, the necessity of fast and secure diagnosis is vital in the medical world. Nowadays, the transmission of visual data is a daily routine and it is necessary to find an efficient way to transmit them over networks. Two main groups of technologies have been developed for this purpose. The first one is based on content protection through encryption. In this group, proper decryption of data requires a key. The second group bases the protection on digital watermarking or data hiding, aimed at secretly embedding a message into the data. In order to not increase the processing time, these two approaches must be combined with the compression stage. Nowadays, the challenge is to perform simultaneously for example image encryption and compression.

International audience; This paper develops a new adaptive scanning methodology for intra frame scalable coding framework based on a subband/wavelet(DWTSB) coding approach for MPEG-4 AVC/H.264 scalable video coding (SVC). It attempts to take advantage of the prior knowledge of the frequencies which are present in different higher frequency subbands. We propose dyadic intra frame coding method with adaptive scan (DWTSB-AS) for each subband as traditional zigzag scan is not suitable for high frequency subbands. Thus, by just modification of the scan order of the intra frame scalable coding framework of H.264, we can get better compression. The proposed algorithm has been theoretically justified and is thoroughly evaluated against the current SVC test model JSVM and DWTSB through extensive coding experiments for scalable coding of intra frame. The simulation results show the proposed scanning algorithm consistently outperforms JSVM and DWTSB in PSNR performance. This results in extra compression for intra frames, along with spatial scalability. Thus Image and video coding applications, traditionally serviced by separate coders, can be efficiently provided by an integrated coding system.

International audience; In this paper, we present a novel method for generating a background model from a sequence of images with moving objects. Our approach is based on non-parametric statistics and robust mode estimation using quasi-continuous histograms (QCH) framework. The proposed method allows the generation of very clear backgrounds without blur effect, and is robust to noise and small variations. Experimental results from real sequences demonstrate the effectiveness of our method.

International audience; Many reversible watermarking solutions have been proposed since 1996: spread-spectrum approaches, circular interpretation on histogram approaches, lossless compression approaches, expansion approaches and histogram approaches. In this paper, we propose a solution whose embedding capacity goes beyond all those reversible schemes. For certain images, the reach payload of the proposed method is over 2 bpp. This solution is an improvement of the Coltuc reversible watermarking approach published in ICIP'2007.

International audience; The diversity of clients in today's network environment compels us to think about solutions that more than satisfy their needs according to their resources. For 3D terrain visualization this translates into two main requirements, namely the scalability and synchronous unification of a disparate data that requires at least two files, the texture image and its corresponding digital elevation model (DEM). In this work the scalability is achieved through the multiresolution discrete wavelet transform (DWT) of the JPEG2000 codec. For the unification of data, a simple DWT-domain spread spectrum (SS) strategy is employed in order to synchronously hide the DEM in the corresponding texture while conserving the JPEG2000 standard file format. Highest possible quality texture is renderable due to the reversible nature of the SS data hiding. As far as DEM quality is concerned, it is ensured through the adaptation of synchronization in embedding that would exclude some highest frequency subbands. To estimate the maximum tolerable error in the DEM according to a given viewpoint, a human visual system (HVS) based psycho-visual analysis is being presented. This analysis is helpful in determining the degree of adaptation in synchronization.

International audience; This tutorial presents the problem of protecting the transmission of medical images. The presented algorithms will be applied to images, videos and 3D objects. The main keywords are compression, encryption, watermarking and data hiding.

International audience; In this paper, the main objective is to establish an existence result of a variational model in image segmentation constrained by a given vector field. In the one dimensional case, we give a discrete version converging in a variational way to the continuous model. We finally describe the numerical analysis of this model with application in image segmentation.

International audience; In this paper a new method is presented for 3D terrain visualization via reversible JPEG2000-based blind data hiding with special focus on data synchronization and scalability. Online real-time 3D terrain visualization involves considerable amount of data. The process is essentially the mapping of the aerial photograph, called texture, onto its corresponding digital elevation model (DEM) implying at least two distinct data inputs. The presence of large disparate data necessitates a compression strategy on one hand and the integration of the DEM and texture into one unit on the other. Whilst the compression must accommodate the scalability requirement originated by the diversity of clients, the unification of data ought to be synchronous. For scalability this paper relies on the multi-resolution nature of the DWT-based JPEG2000 standard whereas the synchronized unification of DEM with the texture is realized by the application of a perceptually transparent data hiding strategy in the DWT domain. The proposed method is blind in the sense that only a secret key, if any, and the size of the original DEM are needed to extract the data from the texture image. We believe that this is one of the pioneering methods to propose scalable embedding of DEM in the texture image. The method is cost effective, in terms of memory and bandwidths, which is an advantage, especially, in real-time environments when quicker transfer of data is required. The results of a 3D visualization simulation effected with our method were encouraging and gave a useful insight to the effectiveness of our method in various bandwidth scenarios.

International audience; For the last five years or so, several 3D data hiding algorithms have been developed. Few of them proposed methods to hide data in 3D objects without changing the position of vertices in the 3D space. In this paper we propose a new approach of data hiding in 3D objects based on the minimum spanning tree (MST). This method is lossless in the sense that the positions of the vertices are unchanged. The method is blind and does not depend on the ordering of the data in the files. Moreover the embedding capacity is significant and can be as high as $2.5 \%$ of the size of the 3D object.

International audience; Small footprint discrete return lidar data have already proved useful for providing information on forest areas. During the last decade, a new generation of airborne laser scanners, called full-waveform (FW) lidar systems, has emerged. They digitize and record the entire backscattered signal of each emitted pulse. Fullwaveform data hold large potentialities. In this study, we investigated the processing of raw full-waveform lidar data for deriving Digital Terrain Model (DTM) and Canopy Height Model (CHM). The main objective of this work was to compare geometric information derived from full-waveform and multi-echo data for various stands. An enhanced peak detection algorithm developed in a previous study was used to extract target positions from full-waveform data on plots under different stand characteristics. The resulting 3D point clouds were compared to the discrete return lidar observations provided by the lidar operator. Ground points were then identified using an original classification algorithm. They were used to derive DTMs which were compared to ground truth. Digital Surface Models were obtained from first echoes and canopy height models were then computed. Detecting weak echoes, when processing full-waveform data, enabled to better describe the canopy shape and to penetrate deeper into forest cover. However DTM was not significantly improved.

International audience; In this paper, we present a novel attack named colorization attack. This attack is specific to color-hiding watermarking schemes. The objective of this work is to demonstrate the feasibility of such an attack and thus to take it into account for the future color-hiding watermarking schemes.

International audience; Day by day, the amount of digital data has been rapidly increasing on the Internet. The size of 3D objects is very large and these objects need fast transmissions. Moreover, 3D data security becomes increasingly important for many applications, e.g., confidential transmission, video surveillance, military and medical applications. In this paper we present two new approaches of 3D object data hiding without changing the position of vertices in the 3D space. The main idea of the two proposed methods is to find and to synchronize particular areas of the 3D objects used to embed the message. The embedding is done by changing the connectivity of edges in the selected areas composed of quadrangles. The first proposed approach of data hiding in 3D objects is based on minimum spanning tree (MST) while the second approach is based on the projection on a secret axis of the quadrangle centers. These methods are lossless in the sense that the positions of the vertices are unchanged. Moreover they are blind and do not depend of the order of the data in the files. These two approaches are very interesting when the 3D objects have been digitalized with high precision.

Extension : 26/04/07 EP1938503 (B1) JP2009513049 (A) US8374340 (B2); La présente invention vise une transmission sécurisée d'un fichier fragmentable en une pluralité de fragments, dans laquelle : (a) on fixe initialement une clé globale de chiffrement, (b) on chiffre le fichier fragmentable en fonction de ladite clé globale pour former un premier fichier crypté ; (c) on chiffre la clé globale au moyen d'un algorithme de chiffrement asymétrique utilisant une clé publique pour obtenir une clé globale chiffrée, (d) on insère des valeurs représentatives de la clé globale chiffrée parmi les fragments dudit fichier crypté pour former un deuxième fichier crypté, et ; (e) on transmet ledit deuxième fichier crypté.

International audience; In this paper, we propose a method to embed the high resolution color information of an image in its corresponding grey-level image. The objective of this work is to allow free access to the grey-level image and give high resolution color image access to secret key owners. This method is made of two steps which are the color image decomposition (in a grey-level image and its associated color information) and the color-information hiding. The main contribution of this paper is to consider the decomposition into 512 colors and to apply a reversible embedding of a bitplane plus a color palette. The resultant watermarked grey-level image is then a 8-bit coded image. With this watermarked image and the extraction secret key, a user is able to rebuild a color image composed of 512 colors. Considering the previous work, the two important consequences of this approach are the improvement of the security and a better color image quality.

International audience; Nowadays, more and more digital images are being sent over computer networks. The works presented in this chapter show how encryption and watermarking algorithms provide security to medical imagery. In order to do this, the images can be encrypted in their source codes in order to apply this functionality at application level. In this way, the encryption and watermarking of images occurs at software level. We can therefore guarantee the protection of a medical image during transmission, and also once this digital data is archived. The subsequent challenge is to ensure that such coding withstands severe treatment such as compression. The quantity of information (entropy) to be sent greatly increases from the original image to the encrypted image. In the case of certain types of medical imagery, large homogenous zones appear. These zones affect the effectiveness of the coding algorithms. Nevertheless, these homogenous zones, useless for any diagnosis, can be safely used for the watermarking of medical images. When a physician receives a visit from a patient, he often requires a specialist opinion before giving a diagnosis. One possible solution is to send images of the patient, along with a specialist report, over a computer network. Nevertheless, computer networks are complex and espionage is a potential risk. We are therefore Chapter written by William PUECH and Gouenou COATRIEUX.

National audience; Nous proposons dans cet article une méthode de protection de Régions d'Intérêt (RI) par insertion de données cachées. Notre méthode comporte une phase de détection automatique des RI (dans le cas présent, les visages), de masquage de ces RI afin d'empêcher une visualisation non-autorisée et enfin, une insertion de données afin de reconstruire la RI sans perte (PSNR infini). La méthode couvre également la phase de décodage de ces vidéos améliorées avec l'extraction du message, la reconstitution de la RI originale et enfin, sa fusion avec la vidéo décodée. Les atouts majeurs de notre méthode sont le faible surpoids entraîné par l'insertion comparé au gain visuel reconstruit dans la région d'intérêt, la protection de la région d'intérêt et la compatibilité avec le standard H.264 utilisé.

International audience; In this paper a novel reversible method for fast and safe image transfer is proposed. The method combines compression, data hiding and partial encryption of images in a single processing step. The proposed approach can embed data into the image according to the message size and partially encrypt the image and the message without changing the original image content. Moreover, during the same process the image is lossless compressed. Nevertheless, the compression rate depends on the upper bound of message size to embed in the image. The main idea is to decompose the original image into two sub-images and to apply various processes to each sub-image in order to gain space and increase the amount of embedded data. The two sub-images are then scrambled and partially encrypted. The most significant characteristic of the proposed method is the utilization of a single procedure to simultaneously perform the compression, the reversible data hiding and the partial encryption rather than using three separate procedures. Our approach reduces then the computational effort and the required computation time. This method is specially suited for medical images where one can associate the patient diagnostic to the concerned medical image for safe transfer purpose.

International audience; In this paper, we propose a method to protect faces in videosurveillance scenes. Our method deletes any visible information of faces in a video and uses a data-hiding technique to embed information in the video that allows further reconstruction of the faces if needed. When the entire information needed for reconstruction is embedded, we obtain infinite PSNR for the face regions.

International audience; In this paper, we propose a method to embed the color information of an high dimension image in its corresponding grey-level image. The objective of this work is to allow a free and rapid access to the grey-level image and give color image access to secret key owners. This method is made of two steps which are the fast high-dimension color image decomposition (in a grey-level image and its associated color information) and the data-hiding. The two main contributions of this paper are the energy function proposed to model the decomposition of the color image and the fast optimization. The optimization of the proposed energy function leads to the achievement of an index image and a color palette. The good properties of that decomposition are an index image which is similar to the luminance of the color image and a color palette which is well suit for the data-hiding. The obtained results confirm the model quality and the real-time property.

National audience; Nous montrons dans ce chapitre comment des algorithmes de cryptage et de marquage permettent la sécurisation des images médicales. De nos jours, de plus en plus d'images numériques sont transférées sur les réseaux informatiques. Aucun système particulier n'a été développé pour le transfert sécurisé d'images. Les travaux présentés dans ce chapitre reposent sur l'idée d'adapter les algorithmes de cryptage et de marquage à l'information image qui est un signal particulier de par sa redondance et de par sa nature bidimensionnelle. Nous proposons de crypter des images au niveau des codages source afin de faire remonter cette fonctionnalité au niveau des couches hautes. De cette manière, les fonctionnalités cryptage et marquage d'images peuvent être insérées au niveau d'un logiciel. De plus, la protection doit être assurée pendant la transmission des images médicales mais aussi pour l'archivage de ces données numériques. Le problème consiste ensuite à faire résister ces chiffrements à des traitements avals comme la compression. En effet, la quantité d'information (entropie) à transmettre augmente fortement entre l'image originale et l'image cryptée. Dans le cas particulier de certains types d'images médicales, des grandes zones homogènes apparaissent. Ces zones perturbent l'efficacité des algorithmes de chiffrement. Paradoxalement ces zones homogènes, peu utilisées pour les diagnostics, peuvent être utilisées en toute sécurité pour le marquage d'images médicales. Aujourd'hui lorsqu'un médecin reçoit un patient il a souvent besoin de l'avis d'un spécialiste avant de prononcer son diagnostic. La solution serait de transmettre les images qu'il a obtenues par liaison informatique. Mais les réseaux informatiques sont complexes et les écoutes illégales nombreuses. Il se pose donc un réel problème quant à la sécurité lors de la transmission de données. Un domaine aussi important et secret que la médecine ne peut donc se permettre de prendre un tel risque sans se protéger. La protection la plus adaptée pour ce type de communication résiderait dans la cryptographie. Beaucoup de techniques de cryptage de texte ont été développées. Depuis l'antiquité, les hommes ont toujours essayé de coder des messages secrets pour se prévenir des oreilles malveillantes. Dans les premières esquisses de cette science du secret, la sécurité résidait dans la confidentialité de l'algorithme qui permettait le chiffrement et le déchiffrement. C'est au fil du temps qu'est apparue progressivement la notion de clef. Aujourd'hui, les systèmes de cryptage reposent sur des algorithmes mis à disposition de tous et c'est la clef, code secret particulier, qui est confidentielle et qui permet de crypter ou de décrypter le message \cite{Kerckhoffs:83}.

International audience; This paper deals with face tracking in difficult conditions of non calibrated camera, strong head motions, thanks to a deformable 3D model. In those conditions, the proposed approach is able to detect and track a face. The novelty is mainly due to a multi-resolution Active Model search which allows to catch strong head motions. Results show an improvement between the single and the multi-resolution technique. Near real-time results are also provided.

International audience; In this paper, we propose a method to embed the color information of an image in a corresponding grey-level image. The objective of this work is to allow free access to the grey-level image and give color image access only if you own a secret key. This method is made of three major steps which are a fast color quantization, an optimized ordering and an adapted data hiding. The principle is to build an index image which is, in the same time, a semantically intelligible grey-level image. In order to obtain this particular index image, which should be robust to data hiding, a layer running algorithm is proceeded to sort the K colors of the palette. The major contributions of this paper are the fast color quantization, the optimized layer running algorithm, the color palette compression and the adapted data hiding.

International audience; The use of aerial photographs, satellite images, scanned maps and digital elevation models necessitates the setting up of strategies for the storage and visualization of these data in an interactive way. In order to obtain a three dimensional visualization it is necessary to map the images, called textures, onto the terrain geometry computed with Digital Elevation Model (DEM). Practically, all of these informations are stored in three different files: DEM, texture and geo-localization of the data. In this paper we propose to save all this information in a single file for the purpose of synchronization. For this, we have developed a wavelet-based embedding method for hiding the data in a color image. The texture images containing hidden DEM data can then be sent from the server to a client in order to effect 3D visualization of terrains. The embedding method is integrable with the JPEG2000 coder to accommodate compression and multi-resolution visualization.

International audience; In this paper, we propose a method to embed the color information of an image in its corresponding grey-level image. The objective of this work is to allow free access to the greylevel image and give color image access to secret key owners. This method is made of two steps which are the color image decomposition (in a grey-level image and its associated color information) and the data-hiding. The main contribution of this papper is the energetic function proposed to model the decomposition of the color image. The optimization of the proposed energetic function leads to the obtention of an index image and a color palette. The good properties of that decomposition are an index image which is similar to the luminance of the color image and a color palette which is well suit for the data-hiding. The obtained results confirm the model quality.

National audience; La visualisation 3D d'un terrain sur des médias de faibles capacités amène à résoudre des problèmes de compression de données et de synchronisation de ces données obtenues, à travers le réseau, à partir de serveurs distants. Le transfert doit aussi pouvoir se faire de façon progressive afin de permettre un affichage sur le média client bien qu'une faible partie des données aura été transmise. Ce travail cherche à résoudre ces problèmes en utilisant le standard JPEG2000 et en particulier le schéma de compression sans perte des données. La compression et la synchronisation des données est assurée grâce à une méthode d'insertion de données cachées permettant de placer toutes les données issues de plusieurs fichiers dans une seule image au format JPEG2000. La scalabilité de la résolution est issue de l'insertion du modèle numérique de terrain, lui même décomposé en ondelettes, dans l'image de texture associée, permettant ainsi un transfert hiérarchique et une synchronisation des données.