Titre : Multiplatform
interactive navigation in large streamed 3D scenes
Encadrants :
Jean-Christophe GONZATO and Pascal GUITTON
Equipe/laboratoire : Projet Iparla / LaBRI-INRIA
Description
:
Recent
advances in acquisition techniques of 3D models (from 3D-scanners to
satellite
radars) allow generating very accurate models. However, this
accuracy
implies to manipulate a large amount of data and thus, specific
strategies
to process them. Moreover, the variety of acquisition techniques
is source
of heterogeneous data structures (meshes, points, implicit
surfaces,
etc.).
Different
solutions were proposed to render and navigate in real-time in
these large
scenes on different hardware from common desktop computers to PC
clusters. However, these techniques are specific and so, not
scalable for
example on handheld devices.
The main challenge of this project is to design remote
adaptive solutions
which could exploit as best as possible the power of any
kind of hardware to
navigate in real-time in a unique dataset. This client/server
approach
implies to take into account the available capabilities (memory,
computing
resources, bandwidth, etc.) of each partner. An adapted data
structure will
have to be proposed in order to allow an efficient streaming
scheme.
This work will be lead using the Elkano framework developed
in the Iparla
Project which allows easy and fast development of new
distributed solutions
for the
remote visualization of 3D models.