Search

Background: Analyzing MR scans of low-grade glioma, with highly accurate segmentation will have an enormous potential in neurosurgery for diagnosis and therapy planning. Low-grade gliomas are mainly distinguished by their infiltrating character and irregular contours, which make the analysis, and therefore the segmentation task, more difficult. Moreover, MRI images show some constraints such as intensity variation and the presence of noise. Methods: To tackle these issues, a novel segmentation method built from the local properties of image is presented in this paper. Phase-based edge detection is estimated locally by the monogenic signal using quadrature filters. This way of detecting edges is, from a theoretical point of view, intensity invariant and responds well to the MR images. To strengthen the tumor detection process, a region-based term is designated locally in order to achieve a local maximum likelihood segmentation of the region of interest. A Gaussian probability distribution is considered to model local images intensities. Results: The proposed model is evaluated using a set of real subjects and synthetic images derived from the Brain Tumor Segmentation challenge –BraTS 2015. In addition, the obtained results are compared to the manual segmentation performed by two experts. Quantitative evaluations are performed using the proposed approach with regard to four related existing methods. Conclusion: The comparison of the proposed method, shows more accurate results than the four existing methods.

Watermarking is used for the protection of intellectual property, data integrity, and data authentication. This paper proposes a novel method for image watermarking based on embedding multiples watermarks in different domains of the image representation (spatial and DCT domains), without any distortion of the watermarked image. In the spatial domain, the processing method is based on study of segmentation by fuzzy c-means clustering method that outputs the zones of watermark embedding and respectively the associated appropriate embedding gain factors. However in the DCT domain a proper choice of the DCT coefficients based on the quantization JPEG table in the middle frequencies band is carried out. Several watermarks were embedded in these two domains in order to take advantage of the spatial domain robustness against different asynchronous attacks, associated to the DCT domain robustness against jpeg compression and some other signal processing distortions. Experimental results show that the proposed method is robust against a large set of synchronous and asynchronous image attacks such as filtering, lossy compression, cropping and rotation attack.

In this paper, we inspect the range and azimuthal defocusing of a buried point scatterer for a low frequency (L or P band) synthetic aperture radar (SAR). Firstly we present the hydraulic electromagnetic and propagation models that allow us to calculate the water content profile in the soil, to deduce the dielectric constant and to simulate the electromagnetic (EM) wave propagation and attenuation. Secondly results of pulse simulation and focusing for several configurations (band polarization, incident angle, point scatterer depth, water content profile) are presented and discussed. The advantages and drawbacks of L and P radars are discussed in the conclusion.

In this paper, we propose a novel digital watermarking scheme in DCT domain based fuzzy inference system and the human visual system to adapt the embedding strength of different blocks. Firstly, the original image is divided into some 8 \times 8 blocks, and then fuzzy inference system according to different textural features and luminance of each block decide adaptively different embedding strengths. The watermark detection adopts correlation technology. Experimental results show that the proposed scheme has good imperceptibility and high robustness to common image processing operators., Comment: 13 pages, 11 figures

L'objectif de ce projet est de construire un atlas indexe d'images et de sequences video en endoscopie digestive et de l'integrer dans un systeme d'aide au diagnostic. Base sur un modele de raisonnement qui differencie l'observation endoscopique et le diagnostic endoscopique, le systeme est compose de 3 elements principaux: une base de connaissance qui reunit la description intensive des classes d'observation et de diagnostic, une base de cas qui rassemble les documents textuels et multimedia bien references et indexes, des interfaces utilisateur. Des tests sur echantillons aleatoires d'observations endoscopiques montrent que la reponse attendue a une requete est unique dans 68 % des cas et premiere d'une reponse multiple dans 28 % des cas. Des tests menes avec des endoscopistes valident le modele et le recensement des indices de la base de connaissance et evaluent les performances du systeme a partir du discours de l'endoscopiste. L'application est accessible par internet et pourrait s'integrer dans un outil d'enseignement de l'endoscopie digestive.

International audience; The goal of any trader is to buy low and sell high, and thus make maximum revenue with minimum risk of loss. In the financial market, prices change on daily bases, leaving traders with confusion about what decision to take, hold, buy, or sell, and when to take it. Many market analysis techniques were introduced to help traders take a winning decision at the right time, one of which was technical analysis. Technical analysis uses indicators to forecast trend and price movements. Therefore this analysis technique aids traders with the decision making process. This analysis technique has shown great success, which made it the resort of most financial traders. However, the efficiency of this type of analysis is affected by many factors, putting into it a great deal of uncertainty, ambiguity and vagueness. In this study, three decision support systems based on a hybrid probabilistic - possibilistic general approach, are proposed and tested on historical prices of the EuroStoXX50, and the CAC40 indices. The following systems take advantage of the statistical claims of probability on historical data, the interpretability and uncertainty handling competences of possibility theory, and the foreseeing abilities of technical indicators, all merged together to arm traders with a reliable daily decision that assures risk - discounted revenue, with a contribution of efficiently taking advantage of multiple indicators.

International audience; In this study, an approach for pixel unmixing based on possibilistic similarity is proposed. This approach, due to the use of possibilistic concepts, enables an important flexibility to integrate both contextual information and a priori knowledge. Possibility distributions are, first, obtained using a priori knowledge given in the form of learning areas delimitated by an expert. These areas serve for the estimation of the probability density functions of different thematic classes. The resulting probability density functions are then transformed into possibility distributions using Dubois-Prade's probability-possibility transformation. Several possibilistic similarity measures were tested in order to improve the discrimination between classes. The pixel unmixing is then performed based on the possibilistic similarity between a local possibility distribution estimated around the considered pixel and the obtained possibility distributions representing the predefined thematic classes in the analyzed image. Accuracy analysis of pixels unmixing demonstrates that the proposed approach represents an efficient estimator of their abundances of the predefined thematic classes and, in turn, higher classification accuracy is achived. Synthetic images are used in order to evaluate the performances of the proposed approach.

International audience; In this study, an iterative contextual approach for images classification is proposed. This approach is based on the use of possibilistic reasoning in order to diffuse the possibilistic knowledge. The use of possibilistic concepts enables an important flexibility for the integration of a context-based additional semantic knowledge source formed by pixels belonging with high certainty to different semantic classes (called possibilistic seeds), into the available knowledge encoded by possibility distributions. The possibilistic seeds extraction and classification process is conducted through the application of a possibilistic contextual rule using the confidence index used as an uncertainty measure. Once possibilistic seeds are extracted and classified, possibility distributions are updated and refined in order to diffuse the possibilistic knowledge. Initial possibility distributions are, first, obtained using a priori knowledge given in the form of learning areas delimitated by an expert. These areas serve for the estimation of the probability distributions of different thematic classes. The resulting probability density functions are then transformed into possibility distributions using Dubois-Prade's probability-possibility transformation. Synthetic and real images are used in order to evaluate the performances of the proposed approach.

International audience; Depuis que les études d'impact des changements climatiques ont montré que le milieu marin pourrait être énormément fragilisé par la disparition de certaines espèces parmi la faune et la flore ainsi que par le vieillissement rapide des infrastructures sous-marines, de nouveaux outils d'observation efficaces et robustes deviennent nécessaires. Dans cet article, l'utilisation de caméras acoustiques comme outil novateur d'acquisition de données sous-marines est proposée en compagnie d'un cadre conceptuel qui permet de mettre en oeuvre une reconstruction tridimensionnelle pertinente et complète de l'environnement sous-marin à partir de séquences d'images acquises par ces caméras acoustiques. Les différentes données et informations extraites, utilisables et utilisées pour élaborer ce travail ainsi que quelques résultats préliminaires sont abordés.

Venous thrombosis (VT) volume assessment, by verifying its risk of progression when anticoagulant or thrombolytic therapies are prescribed, is often necessary to screen life-threatening complications. Commonly, VT volume estimation is done by manual delineation of few contours in the ultrasound (US) image sequence, assuming that the VT has a regular shape and constant radius, thus producing significant errors. This paper presents and evaluates a comprehensive functional approach based on the combination of robust anisotropic diffusion and deformable contours to calculate VT volume in a more accurate manner when applied to freehand 2-D US image sequences. Robust anisotropic filtering reduces image speckle noise without generating incoherent edge discontinuities. Prior knowledge of the VT shape allows initializing the deformable contour, which is then guided by the noise-filtering outcome. Segmented contours are subsequently used to calculate VT volume. The proposed approach is integrated into a system prototype compatible with existing clinical US machines that additionally tracks the acquired images 3-D position and provides a dense Delaunay triangulation required for volume calculation. A predefined robust anisotropic diffusion and deformable contour parameter set enhances the system usability. Experimental results pertinence is assessed by comparison with manual and tetrahedron-based volume computations, using images acquired by two medical experts of eight plastic phantoms and eight in vitro VTs, whose independently measured volume is the reference ground truth. Results show a mean difference between 16 and 35 mm(3) for volumes that vary from 655 to 2826 mm(3). Two in vivo VT volumes are also calculated to illustrate how this approach could be applied in clinical conditions when the real value is unknown. Comparative results for the two experts differ from 1.2% to 10.08% of the smallest estimated value when the image acquisition cadences are similar.

International audience; In this paper, we propose a new Association Rule Mining algorithm for Classification (ARMC). Our algorithm extracts the set of rules, specific to each class, using a fuzzy approach to select the items and does not require the user to provide thresholds. ARMC is experimentaly evaluated and compared to state of the art classification algorithms, namely CBA, PART and RIPPER. Results of experiments on standard UCI benchmarks show that our algorithm outperforms the above mentionned approaches in terms of mean accuracy.

Telemedicine and more largely e-health should be considered as one of the strategic components in the national health care system since this assists the delivery of equitable health-care and the accessibility of specific skills. Through telemedicine patients can get access to medical expertise that may not be able at the patient's site. Advanced technology such as informatics, medical imaging, robotics, voice activated machines, and remote controls have changed the hospital and operating room to accessible zones in all parts of the world and distance has become meaningless in modern times. Telemedicine networks contribute to the improvement of the medical care quality and to the cost effective use of medical resources. It have a strong impact in developing countries since it allows remote parts to get access to medical care and create local knowledge. The Tunisian experience in Telemedicine is significant and has begun on 90th by scientific meetings with local and international links between researchers. Numerous projects and significant networks have been established in order to set up links and connections between Tunisian hospitals and medical centers in one hand and between Tunisian and foreign hospitals in other hand. This article presents the advantages of Telemedicine and its necessity in developing countries. The Tunisian experience is then detailed and discussed.

International audience; In this work, we present a new fusion method that uses fuzzy set theory. This method is applied to the diagnostic system rule bases. It aims at combining all the rule bases into only one rule base and then taking into consideration the characteristics of this base. The fusion method is characterized by a hybrid fusion which combines rule fusion approach with knowledge fusion approach. Knowledge fusion relies on the distortion measure of various bases. This distortion measure is integrated into the rule fusion process in order to generate one rule base for improving the diagnostic system performance. It is defined as the confidence degrees associated to each rule base parameter. The confidence degrees are then integrated into prediction procedure of the new diagnostic system.

ONERA radar RAMSES was recently upgraded with low frequency band (P-band, 435 MHz). In P-band, frequencies penetrate through the ground and through forest canopy. Unfortunately, the formation of the subsurface radar images presents a number of new specific challenges that include algorithm validity, calibration methods, radio-frequency interference, and image focusing and analysis. This paper addresses the aspects of focusing images from an airborne low frequency SAR of buried targets in a lossy and dispersive soil. The hydrological model (Richards' equation) was used to model the multilayer case.

International audience; The exchange of databases between hospitals requires efficient transmission and storage to reduce the cost of health care. This exchange involves a large amount of vital patient information such as patient details, history and measured physiological signals, documents and medical images. This paper presents a technique of interleaving patient information such as text documents and physiological signals with medical images for efficient storage and secure transmission. Our embedding and extracting schemes for digital watermarking of medical images with hidden patient information is proposed as a way to effectively reduce memory requirements, provide protection of information and reduce time and cost of transmission. We present this new method with more details.

International audience; Space-time adaptive processing (STAP) can improve target detectability in a presence of a ground clutter for airborne radar. Ground clutter echo has a wide spectrum as a result of the radar platform (airplane or satellite) motion. To reject clutter echo and preserve target echo, STAP employs antenna array. Simultaneous filtering in both spatial (angle) and frequency domain can improve performance. In this paper we propose a new application of STAP. Assuming an antenna array standing on a sea shore, the objective is to detect targets on the sea surface. Unfortunately sea clutter has different statistical properties compared to airborne clutter. As a consequence, basic STAP algorithm is not optimal in any sense. Future research will face the problem of evaluating performance of STAP applied to sea clutter. We hope to develop a new algorithm that can perform better in the presence of sea clutter.

International audience; Venous thrombosis screening exams use 2D ultrasound images, from which medical experts obtain a rough idea of the thrombosis aspect and infer an approximate volume. Such estimation is essential to follow up the thrombosis evolution. This paper proposes a method to calculate venous thrombosis volume from non-parallel 2D ultrasound images, taking advantage of a priori knowledge about the thrombosis shape. An interactive ellipse fitting contour segmentation extracts the 2D thrombosis contours. Then, a Delaunay triangulation is applied to the set of 2D segmented contours positioned in 3D, and the area that each contour defines, to obtain a global thrombosis 3D surface reconstruction, with a dense triangulation inside the contours. Volume is calculated from the obtained surface and contours triangulation, using a maximum unit normal component approach. Preliminary results obtained on 3 plastic phantoms and 3 in vitro venous thromboses, as well as one in vivo case are presented and discussed. An error rate of volume estimation inferior to 4,5% for the plastic phantoms, and 3,5% for the in vitro venous thromboses was obtained.

When the endoscopist is exploring the digestive cavity, the action of focusing on a particular organ area and of evoking diagnostic hypotheses underlies a complex march of the thought. This mind process has to be analyzed in terms of data fusion and of reasoning in order to lay foundations with a view to a computerized expert system in endoscopy. Provided with evolved functionalities (the similar case retrieval and the diagnostic hypothesis evaluation), a digestive endoscopy atlas can be used as a tool for training and even as a diagnostic aid for "junior" endoscopists. The architecture of such a system is to have its basis in Case Based Reasoning (CBR), organized around two bases, one of endoscopic knowledge and one of case iconography. The retrieval of similar cases can be assimilated by a fusion process with two steps, classification (global method) owing to the knowledge base and, afterwards, measurement (local method) by means of selected cases of the endoscopic case base. Lesion identification and Disease recognition - that, is, the Scene analyze - are the driving forces not only in the Classification stage, but also in the Similarity measure stage. Besides, classification algorithms, tested on simulated endoscopic descriptions, afford results full of promise.

Telemedicine is a growing field and has to face various challenges like its integration to the medical practice and to the health care system, its economic implication and its social impact. This paper will focus on the technological aspects of telemedicine that will be first reviewed. Three main dimensions will be shown to structure the technological space of telemedicine: network, information processing and knowledge engineering. These three components will be briefly discussed. In a second part we will develop three specific issues addressing challenging image and knowledge engineering problems (multimedia medical image management, multimedia medical knowledge bases constitution and management, and case-based reasoning in diagnostic imaging and medical data mining).

Venous thrombosis is a common pathology that creates serious problems in public health. The diagnosis of thromboses, particularly the determination of their relative ages can be efficiently accomplished by ultrasound imaging. This study intends to classify automatically the thrombosis ages by using a predefined learning base that depends on a prior knowledge of physicians. In practice, this learning base is affected by information imperfections of the type ambiguity since physicians cannot give exact thrombosis ages. Thus, the proposed learning base is constructed in a 3-tuple: observation, label, membership value in term of fuzzy logic for each class and not a 2-tuple as in the usual supervised Bayes classification application. By considering this "fuzzy learning base", a method modeling simultaneously the concept of probabilistic uncertainty and ambiguity is proposed. In this approach, the probability for a given observation is considered on the membership value of each class and not on the class itself. At this level, the discussion focuses on two types of applications: the thrombosis age classification and the definition of membership function by using a fuzzy learning base for classification.

The development of a computer-assisted diagnostic system in digestive endoscopy implies to understand the reasoning process of endoscopists. The aim of this study is to validate a reasoning model and a knowledge base previously defined. Eight endoscopists have participated to a diagnostic test including 5 video-sequences and using a "thinking aloud" method. The physician discourse was compared with the knowledge base contents Forty requests, built with indices extracted from endoscopist talks were submitted to the diagnostic research system. The main results show that the endoscopists evoke all the information types defined in the knowledge base. Endoscopists well distinguish the endoscopic findings and diagnoses. They combine data-driven and hypothesis-driven approaches during the diagnostic process. Faced to complex diagnostic problem solving, they promote an analytical approach. The 40 requests have led to the correct finding class, as unique or first response in 25, second or other rank response in 12, and failed to recognize the lesion in 3 cases. These results are consistent with the reasoning model and the knowledge base content They suggest several ways for improving the diagnostic aid system, the research approach as well as the user interface.

International audience; Because of considerable delays introduced by the current diagnostic procedure and lack of dynamic knowledge availability about cervical cancer geographical incidence in Mexico, Information Technology solutions are required to support both diagnostic improvement and epidemiological assessment. We propose an integrated telediagnostic platform and epidemiological information system, based on standard open technologies, capable of epidemiological geospatial information analysis, as a reliable alternative to reduce cervical cancer diagnostic delays, and improve epidemiological knowledge, particularly in rural areas. This paper presents the integrated system architecture, along with its innovative elements like a Specialized Telepathology Electronic Patient Record based on JPEG 2000, an optimized application design that allows synchronous information exchange without dedicated server repositories and an adapted relational database, coupled to a geographical model, to analyze the demographic characteristics of women population being at risk of developing the disease

The computerized and sophisticated version of medical atlases aims at constituting iconographic bases and at developing "intelligent" tools, thus allowing image retrieval according to several modalities. Such atlases are consequently used as assistance tools for physicians, at the same time for medical diagnosis, formation and research. Therefore, the study described is directed at upper digestive endoscopy and its imagery with the aim of conceiving a computer-assisted diagnosis system. It will be articulated around a base of images on which will be supported case-based reasoning), adapted to elucidate a case-in fact, to find the similar ones to it. Moreover, a knowledge base describing endoscopic pathologies is destined to take place of a control unit of coherence for similarities search.

This paper discusses several aspects concerning a diagnosis aid system of esophagus tumors from endosonographic sequences. The question of the aid system role in the clinician diagnosis elaboration is discussed. After a general presentation of the expert reasoning key elements, levels where the practitioner can be assisted are examined. Several features of the system are evoked and hypotheses about its components (case based reasoning) are considered.

Representative of distributed AI, the multi-agent paradigm is of great interest in modeling complex applications: it amounts to a cutting out in autonomous tasks, delegated to entities specialized in an under-field of the system, namely the agents. The difficulty arises in the articulation of these various tasks, i.e. cooperation between agents and management of possible conflicts, in short, the organization and control of the application functioning. On the basis of this general principle, an examination is made of an endoscopic diagnosis aid system with case-based reasoning, integrated in a multi-agent-type architecture worked out in accordance with the "blackboard" model.

Presents a method for the extraction of anatomic structures on ultrasound images using the discrete dynamic contour. The concept of the traditional "snake" is kept while overcoming many of their limitations by using fuzzy images as potential term and by adding constraints on the model. After the model description, results are presented using oesophageal endosonographic images containing tumoral areas.

Image and video sequences take an important place in the medical diagnosis process. In digestive investigation units, endoscopical examinations lead to an important volume of images and video data, which are usually lost as soon as they are produced. The aim of this project is to build a reference data bank of images and video sequences, to organize the content indexing system for implementing and retrieval of specific images, in the field of computer-assisted diagnosis of digestive endoscopical lesions. This paper reports the hypothesis that sustains the project. Indexing of endoscopical images by content refers to 2 different approaches. The operator describes the image in an implicit and symbolic way and joins other levels of information such as consistency of lesion or medical context; the computer scientist uses more elementary and quantitative characteristics. Bringing together the medical symbolic space and the observation space leads to a consistent base of knowledge. The endoscopical image is described as a scene of simple or complex objects each of them characterized by elementary attributes, and these descriptors are listed.

In this paper, we propose a region based image compression approach for medical applications. We use a vector quantization scheme, combined with regions-of-interest. The image to be compressed is first segmented into regions and a separate codebook is used for each specified region. Codeword size and number of codewords by codebooks may be different according to the diagnostic importance of the corresponding image region. This permits to create appropriate codebooks with representative codewords, and to obtain good reconstruction quality in relevant zones, while reinforcing compression in less important regions. The proposed approach is tested on ultrasound esophagus images and is shown to be very promising.

This paper investigates the finite state vector quantization signal compression approach. A new coding scheme is proposed which optimizes the performance of the so-called conditional histogram next-state function design. Optimization is performed by determining for every input block the subcodebook size, that minimizes the expected value of the number of bits in the compressed bit-flow. This is done under the constraint to ensure the same reconstruction quality as that of the full-search VQ. Two different state-correction algorithms are studied. The proposed schema for image compression is tested and is shown to give better results than classical FSVQ approaches.

The coding and transmission of medical images over noisy channels are investigated. The proposed coding scheme is based on vector quantization and the codebook is created by Kohonen's neural approach. In conjunction with the ordered codebook, an appropriate QAM (quadrature amplitude modulation) is used for transmission, which results in eliminating a high ratio of transmission errors. Residual erroneous blocks are eliminated by applying a post-processing method on the received image. Based on the visual similarity between each block and its local context, the impulse block noises are detected and corrected. Using the proposed approach, the reconstructed image quality is acceptable even when the channel is extremely noisy.

Image coding using vector quantization is an interesting approach for image compression. Among the different existing algorithms, Kohonen's self-organizing feature maps (SOFMs) are well suited for designing the codebooks because of their specific properties. Moreover, the use of this method for compression gives, in the same process, basic information on the image content, but to preserve the diagnostic accuracy in echoendoscopic images, it is necessary to take small codewords, generally not greater than subimages of 3 by 3 pixels, which limits the compression rate. We propose to improve the compression rate by using four codebooks, with codewords of different size. Image analysis for coding uses a quadtree scheme. Results are compared with those obtained using the standard JPEG image compression algorithm.

This paper presents a novel, fuzzy-like modification to Hough transforms for circle detection. The technique is applied to ultrasound images of the peri-esophageal region for the localization of the aorta. The results demonstrate promise.

We propose a detection method of esophagus outer wall from endosonographic sequences (composed of separate slices uniformly distributed), which minimizes the information alterations due to the cooperation of different models. The kernel of the proposed solution is based on the use of a stochastic optimization algorithm, fully adapted to our particular case: the goal is to find the optimal contour, which verifies regularity conditions and maximizes a given criteria. Such a method presents the advantage of taking into consideration the entire searched space and thus, avoiding local minimum optimization problems. Moreover, this approach cooperates with a data fusion based processing, which allows a priori knowledge integration with its own inaccuracy. Detection robustness is finally maintained by the use of control agents, which take advantage of adjacent slices. At this level, fuzzy fusion and evidence theory are confronted. All these components are integrated in a coherent and cooperative architecture. Results obtained on real images acquired are very encouraging.