Search

During the last few decades, multimedia security over the cloud has become a major issue. Public-key homomorphism is an efficient approach for data hiding in encrypted images (DHEI). An original image is encrypted using a public key and sent across a network and then processed to embed a secret message directly in the encrypted domain. During the decoding step, a private key is used to obtain a marked reconstructed image, where the secret message can be extracted. In this paper, we propose an efficient method of DHEI based on the Paillier cryptosystem. Using its homomorphic properties, pixel blocks and bits of the message are multiplied in the encrypted domain resulting in an addition in the clear domain. By applying a pre-processing step on the original image before encryption, this addition becomes a least significant bits (LSB) substitution. Experimental results show that using our proposed scheme, we obtain a high payload value ( $1~bpp$ ) without expanding a lot the original image size contrary to current state-of-the-art methods. Indeed, whatever the key size, the expansion rate is always equal to 2. In addition, the original image and the marked reconstructed image are very similar.

International audience; Today, 3D objects are an increasingly popular form of media. It has become necessary to secure them during their transmission or archiving. In this paper, we propose a two tier reversible data hiding method for 3D objects in the encrypted domain. Based on the homomorphic properties of the Paillier cryptosystem, our proposed method embeds a first tier message in the encrypted domain which can be extracted in either the encrypted domain or the clear domain. Indeed, our method produces a marked 3D object which is visually very similar to the original object. It seeks to be format compliant and to preserve the original size of the data, without the need for an auxiliary file. Moreover, large keys are used, rending our method secure for real life applications.

International audience; During the last few decades, the transmission of images over secure networks has exponentially grown. Data security in certain applications such as secure storage, authentication or privacy protection on cloud platforms, require specific strategies for multimedia. Cryptography can be used for this purpose. Indeed, using a secret key, it is possible to make data unreadable in order to secure it. Although encryption approaches are effective to make the original data unreadable, they are also very sensitive to noise. Because of the introduction of noise into an encrypted image during its transmission or storage, the original data cannot be recovered. In this paper, we first describe a new encryption mode called CFB-then-ECB and based on a combination of the CFB mode and the ECB mode for AES encryption. Using this new encryption mode, if one encrypted pixel block is noised, this will result in two incorrectly reconstructed pixel blocks during the decryption (the current and the following pixel blocks). This noise spreading is then exploited in a new proposed approach of noisy encrypted image correction. It contains two main steps involving a classifier to discriminate clear and encrypted pixel blocks. After a direct decryption of a noisy encrypted image, the first step is to identify and localize the pixel blocks that are probably incorrectly decrypted. The second step of our proposed approach is to analyze and correct these pixel blocks. Experimental results show that the proposed method can be used to blindly correct noisy encrypted images, while preserving the image structure without increasing the original data size with additional information.

Today, it has become very easy to manipulate digital images using image processing tools and software such as Adobe Photoshop ( https://www.adobe.com/products/photoshop.html ). Tampering with images by splicing is an operation that consists of cutting-and-pasting an area of an image into another host image. In this paper, we propose to detect and localize such manipulations by analyzing the correlation of the image noise across the three color channels RGB, which is an intrinsic feature of the digital photography acquisition process. More precisely, we propose to detect the border between the background (host image) and the spliced area. Using a sliding window, we detect the blocks that span across the two areas which are characterized by two different color noise correlations. To do this, we propose specific features that are able to highlight these blocks. After the feature extraction, we introduce a learning phase using a Random Forest Classifier. Experimental results, specifically on the Columbia database, show very good results in comparison to other current state of the art methods.

International audience; With the fast development of cloud computing, exchanging JPEG compressed images in a secure way has significantly increased. Data hiding in encrypted images (DHEI) is an effective way to embed additional data directly into the encrypted domain. In recent state-of-the-art methods, almost all DHEI processes focused on uncompressed images. Recently, some schemes using data hiding (DH) in JPEG crypto-compressed images have been designed, but most of them are not fully JPEG format compliant. In this paper, we propose a hierarchical high capacity data hiding (HHCDH) approach for JPEG cryptocompressed images. After encrypting every non-null coefficients, they are processed from low to high frequencies. Sign bits that are specific to them are then substituted by bits of a secret message. During the decoding phase, correlations between neighboring blocks are exploited to hierarchically recover the original sign bit values. According to our experiments, we achieve to obtain a high payload value, while preserving a very good quality of the reconstructed JPEG image. Index Terms-Signal processing in the encrypted domain, data hiding, crypto-compression, JPEG compression.

International audience; Automated image analysis and deep learning tools such as object detection models are being used increasingly by biologists. However, biological datasets often have constraints that are challenging for the use of deep learning. Classes are often imbalanced, similar, or too few for robust learning. In this paper we present a robust method relying on hierarchical classification to perform very small object detection. We illustrate our results on a custom dataset featuring 22 classes of arthropods used to study biodiversity. This dataset shows several constraints that are frequent when using deep learning on biological data with a high class imbalance, some classes learned on only a few training examples and a high similarity between classes. We propose to first perform detection at a super-class level, before performing a detailed classification at a class level. We compare the obtained results with our proposed method to a global detector, trained without hierarchical classification. Our method succeeds in obtaining a mAP of 75 %, while the global detector only achieves a mAP of 48 %. Moreover, our method shows high precision even on classes with the less train examples. Confusions between classes with our method are fewer and are of a lesser impact. We are able achieve a more robust object classification with the use of our proposed method. This method can also enable better control on the model's output which can be particularly valuable when handling ecological, biological or medical data for example.

International audience; Reversible data hiding in encrypted images (RDHEI) can be used as an effective technique to embed additional data directly in the encrypted domain and therefore, without any invasion to privacy. In this way, RDHEI is especially useful for labeling encrypted images in cloud storage. In this paper, we propose a new method of data hiding in encrypted images, which is fully reversible and has a very high payload. All the bit-planes of an image are processed recursively, from the most significant one to the least significant by combining error prediction, reversible adaptation, encryption and embedding. For pixel prediction, the Median Edge Detector, also called LOCO-I and known to be efficient in JPEG-LS compression standard, is used for each bit-plane. Moreover, conversely to current stateof-the-art methods, in our proposed method, there is no preprocessing step to correct incorrectly predicted pixels and no flags to highlight them. Indeed, a reversible adaptation of the bit-planes is performed in order to make it possible to detect and correct all incorrectly predicted pixels during the decoding step. Thanks to the high correlation between pixels in the clear domain, a large part of the bits of an image can be substituted by bits of a secret message. Our experiments show that we can generally embed bits of the secret message until the fourth mostsignificant bit-plane of an image, this allows us to have an average payload value of 2.4586 bpp. Index Terms-Image security, image encryption, reversible data hiding, recursive process, bit-plane prediction, signal processing in the encrypted domain.

International audience; Digital data such as images must be secured during transmission or cloud storage. Image encryption algorithms can be a solution to this problem, but these approaches are very noise sensitive. Because of the introduction of noise, the original image cannot be recovered, even if we know the secret key. In this paper, we propose a new noisy encrypted image correction algorithm containing a convolutional neural network (CNN) training stage and then, two main steps. After a direct decryption of a noisy encrypted image, the first step is to identify and localize the blocks that are probably incorrectly decrypted using a fine-tuned CNN. The second step of our proposed approach is to analyze and correct these blocks. Experimental results show that the proposed method can be used to blindly correct noisy encrypted images, while preserving the image structure and without increasing the original data size with additional information, unlike error correcting codes. Index Terms-Multimedia security, image encryption, image denoising, deep learning, convolutional neural network, signal processing in the encrypted domain.

International audience; The rising popularity of social networks and cloud computing has greatly increased number of JPEG compressed image exchanges. In this context, the security of the transmission channel and/or the cloud storage can be susceptible to privacy leaks. Selective encryption is an efficient tool to mask image content and to protect confidentiality while remaining format-compliant. However, image processing in the encrypted domain is not a trivial task. In this work, we present a JPEG crypto-compression method which allows us to recompress a JPEG crypto-compressed image several times, without any information about the secret key or the original image content. Indeed, using the proposed method in this paper, each recompression can be done directly on the JPEG bitstream by removing the last bit of the code representation of each non-zero coefficient, adapting the entropic code part, and slightly modifying the quantization table. This method is efficient to recompress JPEG crypto-compressed images in terms of ratio compression. Moreover, the decryption of the recompressed image produces an image with a very similar visual quality when compared to the original image, according to the obtained results.

International audience; Today, it has become more frequent and relatively easy to digitize the surface of 3D objects and then to reconstruct a boundary representation (B-Rep), which is a combination of geometric primitives like planes, cylinders, spheres and cones. However, the obtained results suffer from various inaccuracies, mainly due to noisy data. Moreover, the given reconstruction contains only the geometry, without any information of a semantic nature induced by the design process. In this paper, we present an efficient method to extract absolute references which are not explicitly present in an object, such as the orthogonal coordinate system and alignment planes. These are known as datum references, which are used in the CAD modeling process. Then, we show that datum references can be useful through an application for beautification, which consists in the adjustment of primitive parameters to satisfy relations such as parallelism, orthogonality and concentricity. Our objective is to design a fast and automatic method, which is seldom seen in reverse engineering. We show the efficiency and the robustness of our method through experimental results applied on reverse engineered 3D meshes.

Implementing telemedicine solutions has recently become a trend among various research teams at an international level. Telemedicine refers to the use of modern information and communication technologies to meet the needs of citizens, patients, healthcare professionals, and healthcare providers, as well as policy makers. Telemedicine applications are very promising and have great potential; they can play a very important role in service provision by improving access, equity and quality through connecting healthcare facilities and healthcare professionals, and diminishing geographical and physical barriers. However, the transmission and access technologies of medical information raise critical issues that urgently need to be addressed, especially those related to security. Further, medical identity theft is a growing and dangerous crime. Stolen personal information can have a negative financial impact, but stolen medical information cuts to the very core of personal privacy. Medical identity theft already costs billions of dollars each year, and altered medical information can put a person’s health at risk through misdiagnosis, delayed treatment or incorrect prescriptions. Yet, the use of hand devices to store, access, and transmit medical information is outpacing the privacy and security protection on those devices. Therefore, the authenticity of the information and related medical images is of prime concern as they form the basis of inference for diagnostic purposes. In such applications, tamper proofing and guaranteed originality of medical data/information is achieved by embedding some kind of watermark(s) which must be secure and robust against malicious attacks. Robustness and security of medical data/information against attacks is an interesting, challenging area for researchers. Potential researchers are using watermarking and cryptography to disseminate security to the medical data. Further, noted researchers are using watermarking techniques in the field of healthcare to address health data management issues, including source and data authentication, efficient image archiving and retrieval, optimizing bandwidth required to transmit the data, and highlighting diagnostically significant regions.

International audience; Direct observation to study biodiversity can be time consuming, however, other methods often provide indirect measurements and are possibly biased. To solve these problems, images can be a useful tool and ecologists have started to rely more and more on images as a source of data and on automated image analysis. However, the existing methods mostly perform image classification. In this paper we present an efficient method based on object detection to access deeper information the content of an image. Using high resolution images, we built a pipeline to slice the original images, perform detections and later refine these observations. We illustrate the interest of this pipeline by using it on-field images taken in agroforestery banana-coffee systems to study invertebrate communities around the banana pests Cosmopolites sodidus and Metamasius sp. and the interactions between the different animals within this community. Experimental results show that our pipeline reaches 87.8% F1-score and allows us to successfully detect and identify 23 species and ant castes. These 23 species are divided into 7 super-classes, but the ant super-class, that shows more individuals and interactions is described more precisely. We are then able to study the interaction network between different species of this community and identify major predators of banana pests within this ecosystem.

In this research project, we are interested by finding solutions to the problem of image analysis and processing in the encrypted domain. For security reasons, more and more digital data are transferred or stored in the encrypted domain. However, during the transmission or the archiving of encrypted images, it is often necessary to analyze or process them, without knowing the original content or the secret key used during the encryption phase. We propose to work on this problem, by associating theoretical aspects with numerous applications. Our main contributions concern: data hiding in encrypted images, correction of noisy encrypted images, recompression of crypto-compressed images and secret image sharing.

International audience; Social network development raises many issues relating to privacy protection for images. In particular, multi-party privacy protection conflicts can take place when an image is published by only one of its owners. Indeed, privacy settings applied to this image are those of its owner and people on the image are not involved in the process. In this paper, we propose a new hierarchical secret image sharing scheme for social networks in order to answer this problem. Based on the disjunctive multi-level approach of Belenkiy applied to images, this solution ensures user privacy, as shown by our obtained experimental results.

International audience; New issues have arisen in the creation of 3D objects linked to collaboration in 3D workflows. In this kind of usage it may be necessary to allow access and to share a lower quality file of a 3D object for some collaborators. In this paper, we present a Format-Compliant Selective Secret 3D Object Sharing (FCSS3DOS) scheme, which visually protects the content using geometrical distortions by selectively sharing bits of a 3D object's geometry using Shamir's Secret Sharing scheme. The degradation level D is selected before the sharing process and determines how much geometrical distortions are induced into n generated shared 3D objects which are given to n participants. When at least k or more participants agree to combine their own shared 3D objects, then the secret 3D object can be recovered without distortion. The value of k can vary between 2 and n. In experimental results, we analyze the degradation level D that influences the geometrical distortions induced into shared 3D objects.

This paper presents a new method of secret three-dimensional object sharing (S3DOS), which allows sharing of three-dimensional (3-D) objects, while preserving its file format by selectively encrypting a 3-D object in order to sufficiently protect the visual nature of the content. This scheme, named format-compliant selective (k, n) S3DOS , modifies selected bits of the vertices of a 3-D object to protect the visual content by inducing geometrical distortions. These distortions are controlled thanks to the application of a degradation level at the beginning of the sharing process. To reconstruct the secret 3-D object, at least $k$ shared 3-D objects among $n$ have to be combined in order to remove the geometrical distortions and recover the exact original 3-D object. Experimental results are presented and evaluated to showcase the feasibility of the proposed scheme.

International audience; Today, it has become frequent and relatively easy to digitize the surface of 3D objects and then to reconstructa combination of geometric primitives such as planes, cylinders, spheres or cones. However, the given reconstruction contains only geometry, no information of a semantic nature used during the design process is included. In this paper, we present a robust method to recognize specific geometric structures which are not explicitly present in an object, such as features and repetitions. These are known as patterns, which are used in the CAD modeling process. Moreover, the digitization of an object often leads to various inaccuracies, and therefore inaccurate extracted primitives. We also demonstrate how recognized patterns can be useful as an application in beautification, which consists of the adjustment of primitive parameters to satisfy geometrical relations such as parallelism and concentricity. Our objective is to design a fast and automatic method, which is seldom seen in reverse engineering. We show the efficiency and robustness of our method through experimental results applied on reverse engineered 3D meshes.

International audience; In the last few years, with the increasing popularity of cloud computing and the availability of mobile smart devices as well as ubiquitous network connections, more and more users are uploading their personal data to remote servers. However, this can lead to significant security breaches, where confidentiality, integrity and authentication are constantly threatened. To overcome these multiple problems, multimedia data must be secured, for example by means of encryption before transmission and storage. In this survey, we look into the issues involved in handling encrypted multimedia data, and more specifically we focus on reversible data hiding in encrypted images (RDHEI). The aim of this survey is to present the birth and evolution of RDHEI methods over the last 12 years. We first highlight different classes and characteristics of RDHEI, then describe representative RDHEI methods. A comparison table is presented to summarize the key features and achievements of each representative RDHEI method considered in this survey. Finally, we share the future outlook of emerging applications and open research topics relevant to RDHEI for the next 12 years and beyond.

International audience; JPEG is the most used image compression algorithm but block wise DCT compression methods produce artifacts due to coefficient quantization. JPEG decompression can be seen as a reconstruction problem constrained by quantization. In this context, we propose to handle this problem by using interval-valued arithmetic. Our method allows to produce interval-valued image that includes the non-compressed original image. The produced convex set allows to apply constrained Total Variation (TV) reconstruction in order to reduce JPEG artifacts (blocking, grainy effects and high frequency noise). Experiments show visual improvement of JPEG decoding assessed by non-reference quality metric. In addition, the stopping criterion of the TV algorithm is given by this metric which provides evidence about JPEG decompression improvement.

International audience; The authentication of printed documents is an important problem these days. Numerous authentication techniques have been proposed in the relevant literature. One of the most promising solutions uses copy-sensitive graphical codes made of particular patterns. The two Level QR (2LQR) code uses specific textured patterns in order to ensure the sensitivity to duplication process. The authentication test of this code is based on comparing the correlation values between the original and the printed-and-scanned codes with a predetermined threshold. The weakest feature of this technique is the rejection of authentic codes due to the small gap between correlation values for authentic and duplicated code. In this paper , we propose to reduce the number of false-negative results of the authentication test by using a very competitive Super-Resolution (SR) technique. The experimental results show the significant improvement of correlation values when using the images printed-and-scanned once, without increasing the correlation values of duplicated codes. Therefore, the 2LQR code copy sensitivity is not affected by the suggested quality improvement process.

International audience; Many techniques have been presented to protect image content confidentiality. The owner of an image encrypts it using a key and transmits the encrypted image across a network. If the recipient is authorized to access the original content of the image, he can reconstruct it losslessly. However, if during the transmission the encrypted image is noised, some parts of the image can not be deciphered. In order to localize and correct these errors, we propose an approach based on the local Shannon entropy measurement. We first analyze this measure as a function of the block-size. We provide then a full description of our blind error localization and removal process. Experimental results show that the proposed approach, based on local entropy, can be used in practice to correct noisy encrypted images, even with blocks of very small size.

International audience; Images that have been altered and more specifically spliced together have invaded the digital domain due to the ease with which we are able to copy and paste them. To detect such forgeries the digital image processing community is proposing new automatic algorithms designed to help human operators reveal manipulated images. In this paper, we focus on a local detection system, which considers which tampered areas produce local statistical effects that do not impact neighboring areas or the image as a whole. We propose to study how the definition of local blocks, considering their size and overlap, impacts final pixel detection. We also propose new features which are an original way to consider the noise of an image as a colored signal. Indeed, in a non-forged image, there is a high correlation of noise between the three color channels R, G and B. We show that an optimal configuration can be defined and in this case the proposed approach outperforms several previously proposed methods using the same tested dataset, in uncompressed and JPEG modes. Note, in this paper we only focus on feature extraction without using machine learning.

International audience; Unlike full encryption methods, for which the encryption does not preserve the internal structure of files, this paper presents an efficient format compliant selective encryption method for 3D object binary formats. The method encrypts selected bits of the 3D object geometry to visually protect the content without increasing the file size. Geometrical distortions are created in order to partially or fully protect the content , but they do not corrupt 3D scenes for preview purposes. As a function of the degradation level, we can obtain security from visual confidentiality to transparent encryption passing through sufficient encryption providing solutions for different application scenarios. Experimental results and analysis show the efficiency of the proposed method.

International audience; Reversible data hiding in encrypted images (RDHEI) is an effective technique to embed data in the encrypted domain. An original image is encrypted with a secret key and during or after its transmission, it is possible to embed additional information in the encrypted image, without knowing the encryp-tion key or the original content of the image. During the decoding process, the secret message can be extracted and the original image can be reconstructed. In the last few years, RDHEI has started to draw research interest. Indeed, with the development of cloud computing, data privacy has become a real issue. However, none of the existing methods allow us to hide a large amount of information in a reversible manner. In this paper, we propose a new reversible method based on MSB (most significant bit) prediction with a very high capacity. We present two approaches, these are: high capacity reversible data hiding approach with correction of prediction errors and high capacity reversible data hiding approach with embedded prediction errors. With this method, regardless of the approach used, our results are better than those obtained with current state of the art methods, both in terms of reconstructed image quality and embedding capacity.

International audience; In order to preserve cultural heritage, this paper proposes to develop a digital paintings visualization system. Our proposed system mainly consists of extracting regions of interest (ROI) from a digital painting to characterize them. These close-ups are then animated on the basis of the painting characteristics and the artist’s or designer’s aim. In order to obtain interesting results from short video clips, we developed a visual saliency map-based method by using the well-known Itti’s saliency analysis which already proved its efficiency on paintings. The experimental results show the efficiency of our approach and an evaluation based on a Mean Opinion Score validates the proposed method. Eighteen volunters watched for a total of 216 views at 72 video clips build up from 6 paintings, and we found a significant difference between our saliency-based generated videos and random-based generated videos.

International audience; With the development of cloud computing, the growth in information technology has led to serious security issues. For this reason, a lot of multimedia files are stored in encrypted forms. Methods of reversible data hiding in encrypted images (RDHEI) have been designed to provide authentication and integrity in the encrypted domain. The original image is firstly encrypted to ensure confidentiality, by making the content unreadable. A secret message is then embedded in the encrypted image, without the need of the encryption key or any access to the clear content. The challenge lies in finding the best trade-off between embedding capacity and quality of the reconstructed image. In 2008, Puech et al. suggested using the AES algorithm to encrypt an original image and to embed one bit in each block of 16 pixels (payload = 0.0625 bpp) [12]. During the decryption phase, the original image is reconstructed by measuring the standard deviation into each block. In this paper, we propose an improvement to this method, by performing an adaptive local entropy measurement. We can achieve a larger payload without altering the recovered image quality. Our obtained results are very good and better than most of the modern state-of-the-art methods, whilst offering an improved security level with the use of the AES algorithm, defined as the encryption standard by the NIST.

International audience

International audience; Today, it has become more frequent and reasonably easy to digitize the surface of 3D objects. However, the obtained results are often inaccurate and noisy. In this paper, we present an efficient method to analyze a curvature histogram from a digitized 3D surface using a real object. Moreover, we propose to use the curvature histogram analysis for many steps of a reverse engineering process, which can be used to retrieve a CAD model from a digitized one for example. Our objective is to design a fast and fully automated method, which is seldom seen in reverse engineering. Experimental results applied on digitized 3D meshes show the efficiency and the robustness of our proposed method.

International audience; Visual Secret Sharing (VSS) is a type of cryptographic method used to secure digital media such as images by splitting it into n shares. Then, with k or more shares, the secret media can be reconstructed. Without the required number of shares, they are totally useless individually. The purpose of secret sharing methods is to reinforce the cryptographic approach from different points of failure as a single information-container. Instead of Lagrange's interpolation employed by classical methods, an approach based on linear equations and coding theory resolves the classic problems of VSS methods like pixel expansion and information losses at recovery time. In this paper, we propose an efficient (2, n) secret image sharing scheme without loss of information and size increase of shares, with n ≤ 7 based on linear equations encoded into eight bits to encode separately pixels and permutations to maintain high entropy in shares.

International audience; For security reasons, more and more digital data are transferred or stored in encrypted domains. In particular for images , selective format-compliant JPEG encryption methods have been proposed for the last ten years. Since encryption is selective, in order to reduce the processing time and to be format-compliant, it is now necessary to evaluate the confidentiality of these selective crypto-compressed JPEG images. It is known that image quality metrics, such as PSNR or SSIM, give a very low correlation with a mean opinion score (MOS) for low quality images. In this paper, we propose an efficient confidentiality metric based on the visual saliency diffusion. We show experimentally that this metric is well correlated with a MOS and efficient to evaluate the confidentiality of selective crypto-compressed JPEG images.

International audience; The authentication of printed documents is a nowadays challenge. One of the promising solutions for document authentication is the use of copy sensitive graphical codes that can offer data storage and support authentication. Both data decoding and physical authentication are based on comparison between printed-and-scanned samples and original numerical codes. In this paper we want to evaluate different existing correlation measures (Kendall and Spearman) and to propose a new Kendall weighted correlation metric. We propose to evaluate this new method by considering its ability to decode stored messages and to evaluate document authenticity.

International audience; With the development of cloud computing, data privacy has become a major problem. Reversible data hiding in encrypted images (RDHEI) is an effective technique to embed data in the en-crypted domain. Indeed, a lot of methods have been proposed, but none allows a large amount of embedding capacity with a perfect reversibility. In this work, we present a new method of reversible data hiding in encrypted images using MSB (most significant bit) prediction. In order to reconstruct the original image without any errors during the decryption phase, we adapt the to-be-inserted message. Some of the pixels' MSB values are used to highlight the prediction errors and the remaining values are replaced by bits of the secret message. Results show that it is still possible to embed a large message (payload close to 1 bpp).

3D meshes are widely used today in very different domains for examples; game, medical diagnostic, CAD (computed aided design) or more recently 3D printing. In this paper we provide a new data hiding method that has a huge capacity, c p=3×c×(n−1) bits where n is the vertex number of the mesh and c is a non null positive integer. Our proposed method synchronizes vertices along a Hamiltonian path, thus we obtained an ordered list of edges. To do this, we have developed a method based on the displacement of a 3D vertex relative to its father in the path. Its new location is computed with static arithmetic coding (SAC) in order to embed data on each coordinate of a vector defined by an edge. Thus, the proposed method is set as a function of the message in order to control the distortions. Moreover, it allows to set the capacity while achieving a better security. Experimental results show that the method has a high capacity and a low distortion while ensuring security of the hidden message.

International audience; The multimedia traitor tracing field involves the embedding of a collusion secure fingerprint in the host signal to retrieve and prevent any multimedia content fraud. Trendy work aims at providing a tracing system which offers a good protection of the digital content and an efficient tracing process. These challenges depend on reducing the length of the embedded fingerprint and the complexity of the accusation process. Furthermore, addressing these issues becomes more and more relevant in media distribution applications involving an important number of users. In this paper, we propose a secure fingerprinting system based on a two-stage tracing strategy which combines two probabilistic tracing codes: Boneh Shaw with replication scheme and Tardos codes. This strategy is applied to a multilevel hierarchical fingerprint which is then embedded using a DCT-based audio watermarking technique. By taking the advantage of grouping users and applying a weight-based tracing process, the proposed fingerprinting system offers to reduce efficiently the computational costs of the tracing time. It has also a good robustness to different types of attacks. We have carried out different tests to evaluate the performance of the system in terms of robustness and imperceptibility. The experimental results show that the proposed fingerprinting system provides a suitable solution to the the infringement copyright problem in multimedia distribution platforms by reducing significantly the users’ retrieval space and performing good detection results in a reduced time.

International audience; This paper presents one of the first methods allowing the protection of the newly emerging video codec HEVC (High Efficiency Video Coding). Visual protection is achieved through selective encryption (SE) of HEVC-CABAC binstrings in a format compliant manner. The SE approach developed for HEVC is different from that of H.264/AVC in several aspects. Truncated rice code is introduced for binarization of quantized transform coefficients (QTCs) instead of truncated unary code. The encryption space (ES) of binstrings of truncated rice codes is not always dyadic and cannot be represented by an integer number of bits. Hence they cannot be concatenated together to create plaintext for the CFB (Cipher Feedback) mode of AES, which is a self-synchronizing stream cipher for so-called AES-CFB. Another challenge for SE in HEVC concerns the introduction of context, which is adaptive to QTC. This work presents a thorough investigation of HEVC-CABAC from an encryption standpoint. An algorithm is devised for conversion of non-dyadic ES to dyadic, which can be concatenated to form plaintext for AES-CFB. For selectively en-crypted binstrings, the context of truncated rice code for binarization of future syntax elements is guaranteed to remain unchanged. Hence the encrypted bitstream is format-compliant and has exactly the same bit-rate. The proposed technique requires very little processing power and is ideal for playback on hand held devices. The proposed scheme is acceptable for DRM of a wide range of applications , since it protects the contour and motion information, along with texture. Several benchmark video sequences of different resolutions and diverse contents were used for experimental evaluation of the proposed algorithm. A detailed security analysis of the proposed scheme verified the validity of the proposed encryption scheme for content protection in a wide range of applications.

International audience; Perceptual hashing is conventionally used for content identification and authentication. It has applications in database content search, watermarking and image retrieval. Most countermeasures proposed in the literature generally focus on the feature extraction stage to get robust features to authenticate the image, but few studies address the perceptual hashing security achieved by a cryptographic module. When a cryptographic module is employed [1], additional information must be sent to adjust the quantization step. In the perceptual hashing field, we believe that a perceptual hashing system must be robust, secure and generate a final perceptual hash of fixed length. This kind of system should send only the final perceptual hash to the receiver via a secure channel without sending any additional information that would increase the storage space cost and decrease the security. For all of these reasons, in this paper, we propose a theoretical analysis of full perceptual hashing systems that use a quantization module followed by a crypto-compression module. The proposed theoretical analysis is based on a study of the behavior of the extracted features in response to content-preserving/content-changing manipulations that are modeled by Gaussian noise. We then introduce a proposed perceptual hashing scheme based on this theoretical analysis. Finally, several experiments are conducted to validate our approach, by applying Gaussian noise, JPEG compression and low-pass filtering.

International audience; Nowadays it is extremely easy to tamper with images and share them thanks to social media. Identifying the transformation history is imperative to be able to trust these images. We address this problem by using image phylogeny trees, where the root is the image that has been less tampered with and as every generation is obtained from the transformation of its parents, the leaves are the most transformed images. Our method for image phylogeny trees reconstruction is based on a binary decision between two images using JPEG compression artifacts. Experimental results show that when there is no missing image data, the reconstruction is very accurate.

International audience; In the last few years, visual privacy has become a major problem. Because of this, encrypted image processing has received a lot of attention within the scientific and business communities. Data hiding in encrypted images (DHEI) is an effective technique to embed data in the encrypted domain. The owner of an image encrypts it with a secret key and it is still possible to embed additional data without knowing the original content nor the secret key. This secret message can be extracted and the initial image can be recovered in the decoding phase. Recently, DHEI has become an investigative field, but the proposed methods do not allow a large amount of embedding capacity. In this paper, we present a new method based on the MSB (most significant bit) prediction. We suggest to hide one bit per pixel by pre-processing the image to avoid prediction errors and, thereby, to improve the quality of the reconstructed image. We have applied our method to various images and, in every cases, the obtained image is very similar to the original one in terms of PSNR or SSIM.

In multimedia distribution platforms, one of the main challenges is to provide an efficient and accurate tracing process despite the lack of information about the colluders' strategy. Indeed, the original Tardos tracing performance is considered as suboptimal because of its agnostic behavior and conservative accusation regardless the collusion strategy. The Expectation Maximization algorithm has shown to be an efficient solution to estimate the collusion channel and thus to tune the Tardos accusation functions. In this paper, we explore the impact of this algorithm in a group-based tracing scheme to deal with the computational costs and the invariance of the Tardos accusation performance. The tracing scheme we propose benefits from a twofold accusation process. Indeed, in a first time, it is based on a two-level tracing strategy which consists in tracing guilty groups in a first level with the Boneh Shaw tracing code and in retrieving at least one colluder in accused groups with Tardos code in the second level. This strategy has reduced efficiently the decoding process of the Tardos code. The main shift we propose in the second level is to apply the Expectation Maximization algorithm to be tightly tied to collusion yielded by colluders and hence to find the more accurate Tardos accusation functions. The performance of the resulting tracing scheme is evaluated according to different criteria and promising results have been achieved when compared to the existing tracing schemes proposed in the literature.

International audience; In this paper, we propose a new approach for block-based lossless image compression by defining a new semiparametric finite mixture model-based adaptive arithmetic coding. Conventional adaptive arithmetic encoders start encoding a sequence of symbols with a uniform distribution, and they update the frequency of each symbol by incrementing its count after it has been encoded. When encoding an image row by row or block by block, conventional adaptive arithmetic encoders provide the same compression results. In addition, images are normally non-stationary signals, which means that different areas in an image have different probability distributions, so conventional adaptive arithmetic encoders which provide probabilities for the whole image are not very efficient. In the proposed compression scheme, an image is divided into non-overlapping blocks of pixels, which are separately encoded with an appropriate statistical model. Hence, instead of starting to encode each block with a uniform distribution, we propose to start with a probability distribution which is modeled by a semiparametric mixture obtained from the distributions of its neighboring blocks. The semiparametric model parameters are estimated through maximum likelihood using the expectation–maximization algorithm in order to maximize the arithmetic coding efficiency. The results of comparative experiments show that we provide significant improvements over conventional adaptive arithmetic encoders and the state-of-the-art lossless image compression standards.

International audience; The availability of high quality copy machines provides a large amount of printed document counterfeits. Numerous authentication techniques based on security printing, graphical codes, hashing or local hashing have been suggested earlier. In this paper, we propose a novel printed document authentication system based on sensitivity of a new two level QR (2LQR) code to copying process. This 2LQR code contains specific textured patterns, which are sensitive to printing and copying processes. Therefore, it can be used to detect unauthorized document duplication. Experimental results show the efficiency of this 2LQR code for copy detection.

International audience; The quick response (QR) code was designed for storage information and high-speed reading applications. In this paper, we present a new rich QR code that has two storage levels and can be used for document authentication. This new rich QR code, named two-level QR code, has public and private storage levels. The public level is the same as the standard QR code storage level; therefore, it is readable by any classical QR code application. The private level is constructed by replacing the black modules by specific textured patterns. It consists of information encoded using q-ary code with an error correction capacity. This allows us not only to increase the storage capacity of the QR code, but also to distinguish the original document from a copy. This authentication is due to the sensitivity of the used patterns to the print-and-scan (P&S) process. The pattern recognition method that we use to read the second-level information can be used both in a private message sharing and in an authentication scenario. It is based on maximizing the correlation values between P&S degraded patterns and reference patterns. The storage capacity can be significantly improved by increasing the code alphabet q or by increasing the textured pattern size. The experimental results show a perfect restoration of private information. It also highlights the possibility of using this new rich QR code for document authentication.

International audience; This paper presents a new segmentation-based compression scheme for 3D dynamic models. The segmentation process is preformed based on the heat diffusion properties, while exploit- ing temporal and spatial dependencies of the geometry component. 3D affine transforms are used to describe the motion clusters. Weighting of these affine transforms allows to faithfully determine the vertex motions in each cluster. Indeed, both of affine transforms and weighting parameters are determined in order to estimate the positions of vertices in each frame, and consequently compute the temporal prediction errors to be quantized and encoded. In order to improve the coding efficiency, the prediction error quantization is op- timized using a rate control mechanism. Comparative coding tests, for 3D mesh sequences, were conducted to evaluate the coding performance and examine the effectiveness of the proposed coder. The experiments have shown that our codec provides very satisfactory results when compared to the state of the art.

Cryptography based techniques are used to secure confidential data from unauthorized access. These techniques are very good for the security and protection of the data but are very sensitive to noise. A single bit change in encrypted data can have a catastrophic impact on the decrypted data. This paper addresses the problem of removing bit errors in visual data which are encrypted using the AES algorithm in CBC mode (Cipher Block Chaining). We propose a noise removal approach based on the statistical analysis of each block during the decryption process. Three statistical measures are proposed, i.e. the global variance method (GVM), the mean local variance method (MLVM) and the sum of the squared derivative method (SSDM) for error correction. The proposed approach uses local statistics of the visual data and confusion/diffusion properties of the encryption algorithm to remove errors. Experimental results show that the proposed approach gives better results in removing noise and can be used for noise removal in visual data in the encrypted domain. HighlightsThis paper addresses the problem of removing bit errors in visual data which are encrypted using the AES algorithm in CBC mode.Three statistical measures are proposed, i.e. the global variance method (GVM), the mean local variance method (MLVM) and the sum of the squared derivative method (SSDM) for error correction.The proposed approach uses local statistics of the visual data and confusion/diffusion properties of the encryption algorithm to remove errors.Experimental results show that the proposed approach gives better results in removing noise and can be used for noise removal in visual data in the encrypted domain.

High density bar codes are very popular today due to large storage capacity and small code area. But unfortunately, high density versions of QR codes are not being used due to reading problems with most smartphones and flatbed scanner QR code applications. Due to frequent changes and small sizes of the black and white modules, there are reading problems in the QR code binarization and tilt correction processes. The binarization method sets the global or local threshold, and binarizes each pixel separately, that is why they are sensitive to print-and-scan distortion and luminosity. In this paper, we focus on the recognition of high density QR codes. We propose to use the centrality bias of each module to improve the module recognition results. This measure has been used for proposed classification methods and for standard QR code recognition methods. In both cases, the recognition rate was improved, as confirmed by the experimental results. HighlightsRecognition of high density QR codes.Used of the centrality bias.Module recognition results.

International audience; This paper presents an efficient compression algorithm for animated three-dimensional (3D) meshes. First, a segmentation approach is applied to achieve the motion estimation. The main idea is to exploit the temporal coherence of the geometry component by using the heat diffusion properties. The motion of the resulting regions is accurately described by 3D affine transforms. These transforms are computed at the first frame to match the subsequent ones. Second, in order to achieve a good compression performance, an efficient rate control mechanism is proposed to quantize the temporal prediction errors. At this stage, a rate-distortion model is used for quantizing the residual information. Comparative coding tests, for irregular 3D mesh sequences, were conducted to evaluate the coding efficiency of the proposed compression scheme. Simulation results show very promising performances.

International audience; Data hiding has become increasingly important for many applications, like confidential transmission, video surveillance, military and medical applications. In this paper we present a new approach of 3D object data hiding without changing the position of vertices in the 3D space. The main idea of the proposed method is to find and to synchronize particular areas of the 3D objects used to embed the message. The embedding is carried out by changing the connectivity of edges in the selected areas composed of quadruples. The proposed approach of data hiding in 3D objects is based on minimum spanning tree (MST). This method is lossless in the sense that the positions of the vertices are remaining the same before and after embedding. Moreover the method is blind and does not depend of the order of the data in the files. This approach is very interesting when the 3D objects have been digitalized with high precision.

Small footprint full-waveform airborne lidar systems offer large opportunities for improved forest characterization. To take advantage of full-waveform information, this paper presents a new processing method based on the decomposition of waveforms into a sum of parametric functions. The method consists of an enhanced peak detection algorithm combined with advanced echo modelling including Gaussian and generalized Gaussian models. The study focuses on the qualification of the extracted geometric information. Resulting 3D point clouds were compared to the point cloud provided by the operator. 40 to 60% additional points were detected mainly in the lower part of the canopy and in the low vegetation. Their contribution to Digital Terrain Models (DTMs) and Canopy Height Models (CHMs) was then analysed. The quality of DTMs and CHM-based heights was assessed using field measurements on black pine plots under various topographic and stand characteristics. Results showed only slight improvements, up to 5 cm bias ...

International audience; In this paper, we propose a new approach for a block-based lossless image compression using finite mixture models and adaptive arithmetic coding. Conventional arithmetic encoders encode and decode images sample-by-sample in raster scan order. In addition, conventional arithmetic coding models provide the probability distribution for whole source symbols to be compressed or transmitted, including static and adaptive models. However, in the proposed scheme, an image is divided into non-overlapping blocks and then each block is encoded separately by using arithmetic coding. The proposed model provides a probability distribution for each block which is modeled by a mixture of non-parametric distributions by exploiting the high correlation between neighboring blocks. The Expectation-Maximization algorithm is used to find the maximum likelihood mixture parameters in order to maximize the arithmetic coding compression efficiency. The results of comparative experiments show that we provide significant improvements over the state-of-the-art lossless image compression standards and algorithms. In addition, experimental results show that the proposed compression algorithm beats JPEG-LS by 9.7 % when switching between pixel and prediction error domains.

International audience; Medical image processing fuses the image processing technologies in the medical disciplines. Particularly, computed tomography images provide a 3D vision of any part of the human body. These 3D images are generated by CT-Scanner devices. In this paper, we propose an advanced method of CT-Scanner identification from its 3D images. Basically, we analyze the sensor noise in order to identify the source CT-Scanner. For each CT-Scanner, we build three dimension identifiers regarding the three directional axes `X', `Y' and `Z'. The dimension identifier consists of a reference pattern noise and a correlation map. To identify the source CT-Scanner from a tested slice, we compute the correlation between each dimension identifier of each device and this tested slice. The highest correlation value represents an indicator to the source CT-Scanner and the acquisition directional axis. To isolate the pure noise, we use a wavelet based denoising algorithm. Experiments are applied on three different CT-Scanners. 10 3D images are selected from each CT-Scanner, each 3D image is composed of 512 slices. As a result, we are able to identify the acquisition CT-Scanner and the acquisition dimensional axis.