Graduate School of Computer Science and Advanced Technologies

www.epita.fr
Le Kremlin-Bicetre, France

Time filter

Source Type

Discovery of ocean depths concerns every one of us. Recent developments of new technologies, of underwater robotic and of GPS are changing dramatically the way people perceive oceans, not anymore from their surfaces, but through their depths, not in immersion but on-line, not deferred but in real time. This is the objective of Digital Ocean. New technologies are now capable of challenging traditional scuba diving by creating an innovative form of diving, based on science and technology, not on bravery - virtual diving in real time. Navigating freely in the oceanic realm, a dream rooted on mans origins, is now accessible to all. Smart autonomous robots are used to collect underwater digital data that generate sceneries of undersea in 3D interactive imagery. Then, these preprocessed background productions are merged, through mixed reality, with real time 2D videos taken by other distant underwater remote controlled robots teleoperated via the internet, and diffused on-line. As a result, through the web, anyone, anywhere, at any moment - with minimal costs and no risks, will be able to dive virtually in real time, in the most remarkable and secluded sites around the world, unconcerned by sea depths, and get an unmatched quality of images, discover a variety of details, experience the feelings of diving without its constraints and, above all, keeping undisturbed the environment Thus, virtual divers alone or in groups, will be able in few days to discover more than a real diver during all his lifetime. To further enhance virtual diving experience, are integrated in the system a serious video game, a wiki-editing and e-learning platforms, linked with marine specialists. Divers themselves will build their immersed world, keeping virtual sites mirroring real sites, continuously updated and improving in realism. With Digital Ocean, the discovery of the other two thirds of our planets immersed territories will become eventually and irreversibly reachable.


Carlinet E.,Graduate School of Computer Science and Advanced Technologies | Carlinet E.,University Paris Est Creteil | Geraud T.,Graduate School of Computer Science and Advanced Technologies | Geraud T.,University Paris Est Creteil
IEEE Transactions on Image Processing | Year: 2015

The topographic map of a gray-level image, also called tree of shapes, provides a high-level hierarchical representation of the image contents. This representation, invariant to contrast changes and to contrast inversion, has been proved very useful to achieve many image processing and pattern recognition tasks. Its definition relies on the total ordering of pixel values, so this representation does not exist for color images, or more generally, multivariate images. Common workarounds, such as marginal processing, or imposing a total order on data, are not satisfactory and yield many problems. This paper presents a method to build a tree-based representation of multivariate images, which features marginally the same properties of the gray-level tree of shapes. Briefly put, we do not impose an arbitrary ordering on values, but we only rely on the inclusion relationship between shapes in the image definition domain. The interest of having a contrast invariant and self-dual representation of multivariate image is illustrated through several applications (filtering, segmentation, and object recognition) on different types of data: color natural images, document images, satellite hyperspectral imaging, multimodal medical imaging, and videos. © 2015 IEEE.


Wohlmuth M.,Graduate School of Computer Science and Advanced Technologies | Pflaum C.,Graduate School of Computer Science and Advanced Technologies
Physics Procedia | Year: 2010

We present a comprehensive simulation model for high-power 3-level lasers based on a "Dynamic Multimode Analysis" (DMA) [1]. DMA has proven to be a powerful tool to analyze mode competition depending on thermal, spatial, and dynamic effects in the laser resonator and to predict power output, beam quality, and pulse profile for 4-level lasers with M2 < 20 in both q-switched an cw operation. In this work, we generalize the DMA model to 3-level gain media by taking into account the temperature dependence of emission and absorption cross-sections and to highly multimode lasers using generalized mode structures instead of eigenmodes [2]. © 2010 Published by Elsevier B.V.


Hemmer M.,University of Central Florida | Vaupel A.,University of Central Florida | Wohlmuth M.,Graduate School of Computer Science and Advanced Technologies | Richardson M.,University of Central Florida
Applied Physics B: Lasers and Optics | Year: 2012

A novel Nd:YVO 4-based regenerative amplifier system operating in the picosecond regime featuring a volume Bragg grating (VBG) as an intracavity spectral narrowing element is described. This compact amplifier provides pulses with duration of ∼85 ps operating at repetition rates ranging from single shot to 10 kHz. The VBG is used to passively tailor the pulse duration and achieve transform-limited pulses with 50 pm FWHM of spectral linewidth. A Gaussian output beam profile is obtained from the amplifier at all repetition rates. The intracavity VBG also guarantees a high spectral purity by efficiently preventing the build-up of out-of-band ASE. The spectral, spatial and temporal properties of this amplifier make it highly suitable for OPCPA pumping applications. © Springer-Verlag 2011.


Yoruk E.,Bogazici University | Konukoglu E.,Bogazici University | Sankur B.,Bogazici University | Darbon J.,Graduate School of Computer Science and Advanced Technologies
European Signal Processing Conference | Year: 2015

A system has been developed for person identification based on hand images. The images of the left hand of the subjects are captured by a flatbed scanner in an unconstrained pose. The silhouettes of hands are registered to a fixed pose, which involves both rotation and translation of the hand and, separately, of the individual fingers. Independent component features of the hand silhouette images are used for recognition. The classification performance is found to be very satisfactory and it is shown that, at least for groups of one hundred subjects, hand-based recognition becomes a viable and secure access control scheme. © 2004 EUSIPCO.


Feng F.,Graduate School of Computer Science and Advanced Technologies | Pflaum C.,Graduate School of Computer Science and Advanced Technologies
Proceedings of SPIE - The International Society for Optical Engineering | Year: 2012

We derive a new model to simulate passively Q-switched intracavity frequency-doubling solid-state laser. By introducing a nonlinear loss term caused by frequency-doubling crystal into the rate equations ,we can express the effect of second-harmonic generation (SHG). We apply a finite volume discretization on gain medium, saturable absorber and frequency-doubling crystal. "Dynamic Multimode Analysis (DMA)" and several Gaussian modes are utilized. At the end, numerical results of passively Q-switched intracavity frequency-doubling solid-state laser are presented. In the case of large pump radius,chaoic phenomenon can be observed numerically. In order to realize the 3D simulation, we mainly use two technics: One is that common rate equations are extended to a set of 3D multimode rate equations, which calculate photon number for different modes separately. The other is to take into account a finite volume discretization. © 2012 Copyright Society of Photo-Optical Instrumentation Engineers (SPIE).


Muller K.,Graduate School of Computer Science and Advanced Technologies | Schwemmer C.,Graduate School of Computer Science and Advanced Technologies | Hornegger J.,Graduate School of Computer Science and Advanced Technologies | Zheng Y.,Siemens AG | And 6 more authors.
Medical Physics | Year: 2013

Purpose: For interventional cardiac procedures, anatomical and functional information about the cardiac chambers is of major interest. With the technology of angiographic C-arm systems it is possible to reconstruct intraprocedural three-dimensional (3D) images from 2D rotational angiographic projection data (C-arm CT). However, 3D reconstruction of a dynamic object is a fundamental problem in C-arm CT reconstruction. The 2D projections are acquired over a scan time of several seconds, thus the projection data show different states of the heart. A standard FDK reconstruction algorithm would use all acquired data for a filtered backprojection and result in a motion-blurred image. In this approach, a motion compensated reconstruction algorithm requiring knowledge of the 3D heart motion is used. The motion is estimated from a previously presented 3D dynamic surface model. This dynamic surface model results in a sparse motion vector field (MVF) defined at control points. In order to perform a motion compensated reconstruction, a dense motion vector field is required. The dense MVF is generated by interpolation of the sparse MVF. Therefore, the influence of different motion interpolation methods on the reconstructed image quality is evaluated. Methods: Four different interpolation methods, thin-plate splines (TPS), Shepard's method, a smoothed weighting function, and a simple averaging, were evaluated. The reconstruction quality was measured on phantom data, a porcine model as well as on in vivo clinical data sets. As a quality index, the 2D overlap of the forward projected motion compensated reconstructed ventricle and the segmented 2D ventricle blood pool was quantitatively measured with the Dice similarity coefficient and the mean deviation between extracted ventricle contours. For the phantom data set, the normalized root mean square error (nRMSE) and the universal quality index (UQI) were also evaluated in 3D image space. Results: The quantitative evaluation of all experiments showed that TPS interpolation provided the best results. The quantitative results in the phantom experiments showed comparable nRMSE of ≈0.047 ± 0.004 for the TPS and Shepard's method. Only slightly inferior results for the smoothed weighting function and the linear approach were achieved. The UQI resulted in a value of ≈ 99% for all four interpolation methods. On clinical human data sets, the best results were clearly obtained with the TPS interpolation. The mean contour deviation between the TPS reconstruction and the standard FDK reconstruction improved in the three human cases by 1.52, 1.34, and 1.55 mm. The Dice coefficient showed less sensitivity with respect to variations in the ventricle boundary. Conclusions: In this work, the influence of different motion interpolation methods on left ventricle motion compensated tomographic reconstructions was investigated. The best quantitative reconstruction results of a phantom, a porcine, and human clinical data sets were achieved with the TPS approach. In general, the framework of motion estimation using a surface model and motion interpolation to a dense MVF provides the ability for tomographic reconstruction using a motion compensation technique. © 2013 American Association of Physicists in Medicine.


PubMed | New England Eye Center, University of Erlangen Nuremburg, Graduate School of Computer Science and Advanced Technologies and Massachusetts Institute of Technology
Type: Journal Article | Journal: PloS one | Year: 2014

To describe enhanced vitreous imaging for visualization of anatomic features and microstructures within the posterior vitreous and vitreoretinal interface in healthy eyes using swept-source optical coherence tomography (SS-OCT). The study hypothesis was that long-wavelength, high-speed, volumetric SS-OCT with software registration motion correction and vitreous window display or high-dynamic-range (HDR) display improves detection sensitivity of posterior vitreous and vitreoretinal features compared to standard OCT logarithmic scale display.Observational prospective cross-sectional study.Multiple wide-field three-dimensional SS-OCT scans (500500A-scans over 1212 mm2) were obtained using a prototype instrument in 22 eyes of 22 healthy volunteers. A registration motion-correction algorithm was applied to compensate motion and generate a single volumetric dataset. Each volumetric dataset was displayed in three forms: (1) standard logarithmic scale display, enhanced vitreous imaging using (2) vitreous window display and (3) HDR display. Each dataset was reviewed independently by three readers to identify features of the posterior vitreous and vitreoretinal interface. Detection sensitivities for these features were measured for each display method.Features observed included the bursa premacularis (BPM), area of Martegiani, Cloquets/BPM septum, Bergmeister papilla, posterior cortical vitreous (hyaloid) detachment, papillomacular hyaloid detachment, hyaloid attachment to retinal vessel(s), and granular opacities within vitreous cortex, Cloquets canal, and BPM. The detection sensitivity for these features was 75.0% (95%CI: 67.8%-81.1%) using standard logarithmic scale display, 80.6% (95%CI: 73.8%-86.0%) using HDR display, and 91.9% (95%CI: 86.6%-95.2%) using vitreous window display.SS-OCT provides non-invasive, volumetric and measurable in vivo visualization of the anatomic microstructural features of the posterior vitreous and vitreoretinal interface. The vitreous window display provides the highest sensitivity for posterior vitreous and vitreoretinal interface analysis when compared to HDR and standard OCT logarithmic scale display. Enhanced vitreous imaging with SS-OCT may help assess the natural history and treatment response in vitreoretinal interface diseases.


PubMed | Graduate School of Computer Science and Advanced Technologies
Type: Journal Article | Journal: Physics in medicine and biology | Year: 2014

We present a software, called CoroEval, for the evaluation of 3D coronary vessel reconstructions from clinical data. It runs on multiple operating systems and is designed to be independent of the imaging modality used. At this point, its purpose is the comparison of reconstruction algorithms or acquisition protocols, not the clinical diagnosis. Implemented metrics are vessel sharpness and diameter. All measurements are taken from the raw intensity data to be independent of display windowing functions. The user can either import a vessel centreline segmentation from other software, or perform a manual segmentation in CoroEval. An automated segmentation correction algorithm is provided to improve non-perfect centrelines. With default settings, measurements are taken at 1 mm intervals along the vessel centreline and from 10 different angles at each measurement point. This allows for outlier detection and noise-robust measurements without the burden and subjectivity a manual measurement process would incur. Graphical measurement results can be directly exported to vector or bitmap graphics for integration into scientific publications. Centreline and lumen segmentations can be exported as point clouds and in various mesh formats. We evaluated the diameter measurement process using three phantom datasets. An average deviation of 0.03 0.03 mm was found. The software is available in binary and source code form at http://www5.cs.fau.de/CoroEval/.


PubMed | Ludwig Maximilians University of Munich, New England Eye Center, University Erlangen Nuremburg, Graduate School of Computer Science and Advanced Technologies and Massachusetts Institute of Technology
Type: Journal Article | Journal: Ophthalmology | Year: 2014

To characterize en face features of the retinal pigment epithelium (RPE) and choroid in eyes with chronic central serous chorioretinopathy (CSCR) using a high-speed, enhanced-depth swept-source optical coherence tomography (SS-OCT) prototype.Consecutive patients with chronic CSCR were prospectively examined with SS-OCT.Fifteen eyes of 13 patients.Three-dimensional 66 mm macular cube raster scans were obtained with SS-OCT operating at 1050 nm wavelength and 100000 A-lines/sec with 6 m axial resolution. Segmentation of the RPE generated a reference surface; en face SS-OCT images of the RPE and choroid were extracted at varying depths every 3.5 m (1 pixel). Abnormal features were characterized by systematic analysis of multimodal fundus imaging, including color photographs, fundus autofluorescence, fluorescein angiography, and indocyanine-green angiography (ICGA).En face SS-OCT morphology of the RPE and individual choroidal layers.En face SS-OCT imaging at the RPE level revealed absence of signal corresponding to RPE detachment or RPE loss in 15 of 15 (100%) eyes. En face SS-OCT imaging at the choriocapillaris level showed focally enlarged vessels in 8 of 15 eyes (53%). At the level of Sattlers layer, en face SS-OCT documented focal choroidal dilation in 8 of 15 eyes (53%) and diffuse choroidal dilation in 7 of 15 eyes (47%). At the level of Hallers layer, these same features were observed in 3 of 15 eyes (20%) and 12 of 15 eyes (80%), respectively. In all affected eyes, these choroidal vascular abnormalities were seen just below areas of RPE abnormalities. In 2 eyes with secondary choroidal neovascularization (CNV), distinct en face SS-OCT features corresponded to the neovascular lesions.High-speed, enhanced-depth SS-OCT at 1050 nm wavelength enables the visualization of pathologic features of the RPE and choroid in eyes with chronic CSCR not usually appreciated with standard spectral domain (SD) OCT. En face SS-OCT imaging seems to be a useful tool in the identification of CNV without the use of angiography. This in vivo documentation of the RPE and choroidal vasculature at variable depths may help elucidate the pathophysiology of disease and can contribute to the diagnosis and management of chronic CSCR.

Loading Graduate School of Computer Science and Advanced Technologies collaborators
Loading Graduate School of Computer Science and Advanced Technologies collaborators