|
|
Version
française
I3D : 3D Interaction
i3D is a recent research group dedicated to one of the main challenges of virtual reality, interaction. Many models and algorithms used in virtual reality can be borrowed from more traditional workstation applications (visualization, simulation,..). However, one of the main pluses, hence primary challenges, of virtual reality is multisensorial, immersive interaction.
i3D concentrates on 3D interaction primarily for virtual environments.
Research is being carried out in particular on large screen immersive visualization
devices with haptic feedback.
"Our long term objective is to contribute to making interaction with virtual
worlds as simple, intuitive and comfortable as with the real world," says
Sabine Coquillart. In this research, i3D comes in on several complementary
levels: (1) study and design of interaction paradigms or metaphors, (2)
haptic interaction, (3) interaction techniques or peripheral evaluation
and human factors.
Since its creation, i3D has selected a virtual reality configuration that is the basis of most of its work. The configuration is a virtual workbench. This choice was motivated by the very strong potential of this configuration in terms of interaction, due to the direct handling of the virtual model, among other factors.
One of the goals of the i3D group is to enrich this configuration, and more generally all virtual reality configurations, with additional interaction and sensory feedback paradigms, such as force feedback.
The i3D teams have already developed specific tools such as:
i3D is also working on more fundamental aspects such as psychophysical evaluation and non active haptic feedback modes such as pseudo-haptic feedback.
The applications envisioned are the visualization, handling and interactive simulation of complex data in such varied application fields as engineering, geoscience, chemistry and biology, among others.
i3D has always maintained productive collaborations with many industrial or academic partners in France and abroad, such as:
Sabine Coquillart
INRIA Rocquencourt
Tél. : + 33 1 39 63 54 23
INRIA Rhône-Alpes
Tél. : + 33 4 76 61 52 65
SIAMES : Computer graphics, animation, modeling and simulation
The team focuses on two main topics, movement simulation and virtual reality.
The goal is to devise "physical" models whose characteristics are given
by the laws of mechanics, for example for the animation of vehicles or the
displacement of virtual objects, or of biomechanics for living being animation.
One of the first applications concerns the simulation of the behavior of
autonomous entities - vehicles or humans - in city surroundings.
Researchers of INRIA Rennes - IRISA are working in collaboration with the Department of Biomechanics and Sports Physiology of the University of Rennes 2, on the analysis of human movements and the measurement of sports gestures for performance evaluation. Transfer to industry was initiated with such partners as Infogrames, Realviz and the Leti in the framework of RNTL projects (1) in the field of motion capture processing of human movements for video games. The "Movement" RNTL project intends to built a multi-sensor motion capture module integrating constraints and biomechanical models.
Concerning virtual reality, project Siames studies new interaction techniques such as force-feedback. A crucial point of this activity concerns the Open Source availability on the Web of a virtual reality software platform OpenMASK . This cross-platform tool integrates immersive visualization, interaction and collaboration capabilities. Many academic and industrial collaborations of the team are based on OpenMASK. In particular, the PEFR-RV RNTL project intends to factor the research and development on virtual reality for future design offices into different domains of application.
Industry transfer initiatives are ongoing in particular with Renault on the theme of force-feedback and collaborative work in virtual reality (PERF-RV), France Telecom R&D in Rennes in the control of virtual humanoids, Giat-Industry for maintenance training in virtual reality and Kineo and Daesign in the field of interactive fiction for virtual humanoids.
(1) National Research Network on Software Technology set up by the government in 1999
Bruno Arnaldi
INRIA Rennes - IRISA
Tél. : + 33 2 99 84 72 61
REVES : Sound-enhanced virtual rendering and environments
Project REVES was founded in 2000 with the goal of developing new algorithms to improve and accelerate the computer image and spatial sound generation process. The rendering algorithms apply to virtual or augmented (a mix of virtual and real) environments. The fields of application are virtual cultural heritage, construction and city planning, video games and audiovisual.
The originality of project REVES is in the coherent study and modeling of sound and image for virtual environments. "It is now widely recognized that adding spatial sound to convincing 3D graphics significantly improves the feeling of presence and immersion in virtual environments," says Nicolas Tsingos, an INRIA researcher of project REVES specializing in spatial sound for virtual environments.
REVES is contributing to audio-visual virtual environment rendering for which sound is an integral part of the experience. In this regard, REVES develops new real time algorithms for 3D sound generation, multimodal perceptual criteria based rendering optimization, artificial reverberation and the modeling of the acoustics of virtual venues, as well as combined management of graphical and audio hardware resources.
REVES has developed a common software platform called REVESAPI for sound enhanced rendering and virtual environments. Several modules are being integrated into this common platform:
REVES is also in contact with several companies that are interested in various aspects of this software.
Project REVES is strongly involved in virtual and augmented environments in the framework of collaborations and financed programs:
INRIA Sophia Antipolis has acquired in April 2000 a Barco Baron virtual workbench, under the responsibility of project REVES. This device is driven by a PC with a standard graphics card, which makes it a lot more affordable than previous installations based on top-of-the-line computers.
George Drettakis
INRIA Sophia Antipolis
Tel : +33.4.92.38.50.32
EPIDAURE : Medical imaging and robotics
Project EPIDAURE was launched in October 1992. It now has some twenty researchers. The project goal is to design and develop new tools for the analysis of multidimensional, multimodal medical images (scan, magnetic resonance imaging, ultrasound, nuclear medicine and so forth) to improve diagnoses and therapies, especially when guidance by images is possible (video-surgery, interventional radiology, radiotherapy, etc.).
Based on the analysis of medical images, it is possible to construct geometrical and biomechanical representations that allow for a better visual or even haptic (force feedback) interaction of doctors with images. It is for example possible to make visual models of the organs of a patient that can be used in two different ways:
Surgical simulation
The objective of surgical simulation is to contribute to the training of
surgeons in complex surgical techniques such as mini-invasive surgery. There
is a parallel to be made with the flight simulators used in aircraft pilot
training.
Using a simulator, a surgeon can interactively handle organ models obtained
by the finite element method, in order to rehearse an operation without
any risks for the patient. Since 1997, project Epidaure has been developing
an hepatic surgery simulator that makes it possible to simulate the resection
of one or several anatomical segments of the liver using two force feedback
systems. A concerted research initiative coordinated by project Epidaure
from 1998 to 2000, including several INRIA projects and IRCAD (Institute
for Research on Digestive Tract Cancers), resulted in significant scientific
advances in this field.
Augmented reality aided surgery
During a surgical operation, a video image of the patient is superimposed with virtual images of his or her anatomy and pathologies. The virtual images are obtained by processing several medical images taken before the operation. The objective of such tools is to help the surgeon better locate in space the anatomical and pathological structures that cannot be observed directly. One of the difficulties in developing these tools resides in the merging and gating of preoperation images with the data observed during the operation.
Nicholas Ayache
INRIA Sophia Antipolis
Tél. : + 33 4 92 38 76 61
ISA : Image, synthesis, analysis
Project ISA is carrying out research in two complementary fields:
The software written concern many fields: architecture, Earth sciences, design and engineering, medical imaging. It is generally developed in collaboration with industry. In addition, three startups, T-surf, Neoxy and VSP-Technology, were founded by project members.
The main contributions of ISA to virtual reality have to do with improving realism and the precision of simulation methods, coupled with immersive visualization methods. The development of new algorithms has applications in the fields of architectural representation and scientific visualization that call upon virtual environments. One problem for example is to devise solutions for lighting by radiosity that are at the same time precise and fast, in order to create realistic visual surroundings for virtual reality applications in architecture. VSP-Technology markets photorealistic simulation solutions coupled with real time visualization engines based on these new approaches.
In the field of geoscience, the ISA team is working on volumic rendering of large scale data and on the interface of the gOcad software (see below) with virtual reality to give users a feeling of immersion in the subsoil. Oil industry specialists are thus able to visualize oil deposits based on seismic data. Immersive environments are a means of visually entering inside the subsoil, virtually dug galleries and simulate oil well drilling. The gOcad software (Geological Object Computer Aided Design), developed at the Geology School of Nancy (part of the Lorraine National Polytechnic Institute-INPL) in collaboration with researchers from project ISA, makes it possible to model and visualize geologic strata based on data acquired via various geophysics techniques. gOcad is marketed by T-Surf, a startup stemming from the gOcad Consortium and based in Nancy and Houston, Texas.
Other work concerns augmented reality methods and the development of robust automatic tools. The goal is to automatically inlay a movie with virtual objects, taking into account the relative positions of real and inserted objects in order to deal with occultation issues in particular. The computation of camera movements and occultation solving are obtained based on a photometric analysis of the movie. Augmented reality is also used in a medical application in angiography, in collaboration with the Nancy teaching hospital.
The PLASMA cooperative research initiative focuses its work on fusion plasma physics. The stakes of this research (a potential energy source for the future) is commensurate with the difficulties of realistically modeling and visualizing the data obtained. The use of immersive environments has imposed itself as an efficient and unavoidable solution.
Jean-Claude Paul
INRIA Lorraine - Loria
Tél. : + 33 3 83 59 20 77
SHARP : Automatic programming and decision systems in robotics
Project SHARP’s research activity is centered of the study of problems connected to physical interactions in robotics, modeling and automatic generation of movement both in the real and virtual worlds. The former case concerns robots, the latter dynamic simulation.
SHARP is developing efficient geometrical algorithms to process interactions between rigid or deformable objects. The project also designs and implements a modeling and simulation process that is capable of taking into account real world data, and develops computer methods allowing for an haptic interaction with a robot.
The research approach emphasizes physical realism and real time constraints in a dynamic virtual scene. The main challenge is posed by the complexity of the objects to be modeled, their heterogeneity, the nonlinearities and the real time aspect.
The applications especially intended by the team are in medical robotics and in physical movement in a virtual world. They touch on surgical simulation, the creation of deformable models adapted to the simulation of certain bodies (Pug prize), the development of collision computation algorithms (Icarcv prize), the creation of autonomous reactive virtual characters and the development of tools for video games.
Christian Laugier
INRIA Rhône-Alpes
Tél. : + 33 4 76 61 52 22
ALCOVE : Working and collaborating on complex virtual objects
The ALCOVE team was set up under the direction of Christophe Chaillou based on a few simple assessments: the computer tools for exchange and cooperation are still basic and not very intuitive, 3D environments and objects are being increasingly used but their interaction capabilities with users remain rudimentary.
ALCOVE is developing new concepts - interaction metaphors and autonomous objects - to create new tools to work on complex virtual objects in group.
Faced with the heavy technological aspects developed in certain research activities (immersive rooms, etc.), ALCOVE is exploring a new course by relying on the development of innovative tools that make it possible for a group to represent and handle virtual objects, with as natural as possible interactions.
This work underlies applications that range from gesture and concept learning to scenario preparation and object design. The five year objective is to propose a work terminal composed of visual and gestual interfaces, simulation and network, to let a distributed surgical team - each participant in front of a terminal - to make a medical diagnosis for a 3D reconstructed patient.
The research team already boasts various achievements, through diverse collaborations with France Telecom (Spin 3D platform).
Christophe Chaillou
INRIA Futurs
Tél. : +33 3 20 43 47 20
IPARLA : Visualizing and handling complex data on mobile communicating terminals
IPARLA is interested in computer image production and possible interactions with a complex virtual environment on a PDA or a mobile phone on a network (GPRS, UMTS, Wi-Fi). This notion of complexity concerns the quantity of data (urban environment or terrain) as well as its nature (non perceptible scientific information, financial information).
Faced with the numerous constraints due to the context (small screen, small processor, no local storage), IPARLA researchers intend to consider algorithmic chains that are different from the classical 3D algorithmics. The research directions considered are based on non photorealistic rendering, pointwise models, multiresolution approaches, client-server architectures and taking into account the perception phenomena in the human brain.
Potential applications are focused on guidance in city or country surroundings (pedestrians, cyclists, car drivers), better directional location (touristic site vists, operation on an industrial site), surveillance (medicine, scientific experiments), complex process running and games.
Collaborations are ongoing with BeTomorrow and France Télécom R&D.
Pascal Guitton
INRIA Futurs
Tél. : + 33 5 56 84 69 18
