The position will take place in Strasbourg within the MIMESIS research team from Inria, the French National Institute for Computer Science and Applied Mathematics.
MIMESIS research team focuses on its global objective which is to create a synergy between clinicians and scientists in order to develop new technologies capable of redefining healthcare, with a strong emphasis on clinical translation. In our case we joined IHU institute for this purpose.
The scientific objectives of our team, MIMESIS, are related to this ambitious objective. Over the past years we have developed new approaches supporting advanced simulations in the context of simulations for training. We now propose to focus our research on the use of real-time simulations for per-operative guidance. The underlying objectives include numerical techniques for real-time computations and data-driven simulation dedicated to patient-specific modeling. This last topic is a transversal research theme and raises several open problems, ranging from non-rigid registration to augmented reality.
Although our team is attached to the Inria Nancy research center, our offices are located inside the IHU Strasbourg.
For more information on the research team, do not hesitate to visit our website at http://mimesis.inria.fr.
During this internship, the candidate will be brought to develop a soft body simulation model. She (he) will have the opportunity to learn advanced animation techniques based on physical dynamics. Surrounded by a young team of dynamic researchers, her (his) work will be integrated into the team’s multi-physics simulation software: SOFA Framework. At the end of the internship, the candidate will have learned the basics of physics-based animation of deformable objects. She (he) will also have been introduced to scientific publications reading in the field of computer animation and simulation in a surgical environment.
This year, we have two internship positions available.
Position-based dynamics In this project, the field of animation based on position constraints, often call position-based dynamics, is explored. Given a cloud of points filling the interior of a simulated soft object, constraints are added to those points in order to force them to regain a steady state (the rest form of the object) after being deformed by some external forces (gravity, collisions, etc.). Those constraints need to be solved at each time step of the simulation. An equilibrium state should be reached after a set of solver iterations, resulting in a deformed object that exactly balances out the external force energy (which try to deform the object) against the internal elastic energy (which try to regain the undeformed state of the object).
Immersed boundary method In this project, a new method of 3D space integration is explored. Given a 3D surface of a simulated soft object (consisting of a set of triangles or quads), and given a set of regular cubes surrounding the object, we wish to compute the set of triangles or quads that intersect the cubes. If a triangle or quad intersects a cube, but isn’t entirely contained inside that cube, this surface element must be split into smaller triangles or quads that will be entirely contained inside the cube. The candidate will have to implement a fast and stable method that compute this new surface where every triangle or quad are contained inside the cubes. The implemented method must allow for non-convex surface, and must be fast enough for real-time applications (optimization with GPU and CPU vector operations can be used).
We are looking for a candidate with a particular interest in numerical simulation of physical phenomena and the development of real-time animation software. Since the trainee will have to use a variety of mathematical tools, we will prioritize candidates who already have experience in numerical computing, or demonstrate a distinctive curiosity for this area of research. The applicant must be self-taught and ready for scientific challenges.