|
|
| Front Page | INédit | Contents of n°64 | Conferences and Events | Press |
|
 |
© Visages/U746 project-team |
The VISAGES U746 project-team has just patented an image readjustment procedure that facilitates neurosurgical operations. The neurosurgeon is assisted by a navigation system that works like GPS to show the trajectory of the surgeon's instruments on an MRI of the patient's brain. This image, recorded before the operation, is static and cannot be used to take into account the way tissue moves and distorts, sometimes substantially, during an operation. This technology patented by Visages, which researchers have been working on for several years (see Inédit No. 53), compensates for these topological changes by combining the MRI with ultrasound data obtained during the operation. Regions that appear brighter on the ultrasound represent tissue folds and are superimposed over the MRI in real time, using mainly measurements of local curvatures as a basis. The surgeon can use this information to assess the movements and distortions in soft tissue, and even to reverse the change in order to improve his or her bearings.
Christian Barillot
CNRS researcher
VISAGES, INRIA project-team shared with the CNRS and the University of Rennes 1 under Irisa (UMR 6074) and Inserm as an Inserm unit (U746)
Tel.: + 33 2 99 84 75 05
 |
© Stéphane Redon/projet Nano-D |
Today, it is possible to modify complex molecules interactively on a common computer using the adaptive simulation software developed by Stéphane Redon, of the INRIA NANO-D team, and his colleagues at the Grenoble CEA. AMDToolkit can be used to adapt to reduced computing capacity by decreasing the precision of the representation and focusing efforts on the important effects. This means that, as needed, users can adjust the level of precision in the simulation while still producing a useful approximation of the molecule's movements.
To achieve this feat, AMDToolkit combines a hierarchical representation of the molecule with a clever system for identifying its most rigid components. The molecule is described as an assembly of two smaller pieces, and so on down to the atom level. This type of representation can be used to easily identify the groups of atoms involved in the tiniest movements. It does so by calculating the aggregated magnitude of mobility for sets of nodes (sum of the squares of torsion angle acceleration figures), where the lowest values correspond to the least mobile sets. These last sets are considered rigid groups, and the positions of atoms within these groups in relation to one another is not recalculated. Only the most mobile links, which vary in number depending on the desired precision, are recalculated; this saves time, especially in cases where little precision is required. The mobile and rigid areas in the molecule are automatically modified to adapt to the user's actions and to changes in the atomic environment.
This technology, originally developed for robotics, is useful not only for simulations in biotechnology but also for virtual prototyping on an atomic scale. A patent application has been filed for the adaptive method for calculating inter-atomic forces developed at INRIA. Researchers now hope to work on a pre-industrial version of the AMDToolkit prototype.
Stéphane Redon
NANO-D team, INRIA Grenoble - Rhône-Alpes
Tel.: + 33 04 76 61 55 69
http://interstices.info/proto-nano
