Textures, Rendering and Visualization

Staff

Researcher – 2,25 ETPR : Rémi Allègre (MC), Jean-Michel Dischler (PR), Jonathan Sarton (MC), Basile Sauvage (MC HDR) Engineer : Sylvain Thery (IR) PhD students : Vinojan Rajandiran (2023-2026), Romain Fournier (2022-2025), Erwan Duhamel (2021-2024), Guillaume Baldi (2020-2024), Charline Grenier (2020-2024)

Objectives

A large part of the activities in this research team focuses on procedural texture generation. Textures are images encoding light behavior (color, brightness, etc.). They are applied to the surface of 3D objects in a virtual scene to "render" the scene, i.e. to calculate computer-generated images. Procedural generation consists in developing computing processes with low storage costs, which can be computed in parallel (on graphics cards), and on a multi-scale basis (in constant time, whatever the scale of observation). This work is at the crossroads of computer graphics, simulation (of light transport) and signal processing.

1. One line of research concerns the rendering of complex materials (i.e. glittering or iridescent). We will continue our collaboration with the Karlsruhe Institut of Technology (KIT).

2. Another topic concerns the interaction between the appearance and shape of objects. In particular, the generation of textures directly on the surface of 3D objects provides an opportunity to collaborate internally within the IGG team, with colleagues specializing in geometry.

3. A third topic concerns the resolution of inverse problems based on our procedural generation methods, i.e. the estimation of parameters from input images. The aim is to develop our skills in the field of AI (machine learning), while continuing the collaboration initiated with colleagues in the Data and Knowledge Sciences team.

4. Finally, we develop activities linked to data augmentation for training deep learning models. One applications is in medical imaging, in particular histopathological imaging, where the ability to generate artificial images resembling real images of biological tissue is a crucial issue.

In addition, we develop new activities in scientific visualization, particularly in high-performance computing environments. Here, we reinforce existing collaborations with the IRMA mathematics laboratory, for the visualization of unstructured meshes resulting from large-scale numerical simulations. The aim is to address aspects of in-situ visualization and volume rendering of complex (topology, geometry, multi-varied, time-varying, non-linear), high-dimensional meshes.

Projets

ANR LUM-Vis Visualization is a tool that is getting more and more widespread, and has become essential today in almost every scientific fields. It is embedded in most experimentation workflows, at different levels. Visualization tools are used to analyze datasets or extract information from them, to guide phenomenon modeling, to validate or invalidate models or as a tool for evaluating experimental results. We intend to overcome classical direct volume visualisation of AMR data issues to establish GPU- based modern volume ray-casting as a reliable tool for in-situ visualization of large-scale simulation that produces large and complex volume data. PhD thesis of Vinojan Rajandiran.

ANR ArtIC : Artificial Intelligence for Care. PhD thesis of Guillaume Baldi.

ANR-DFG ReProcTex Rendering procedural textures for huge digital world is a joint research project between two research groups in computer graphics, at University of Strasbourg (France) and at Karlsruhe Institute of Technology (Germany). Current virtual worlds are huge. See for example virtual film sets, cultural heritage visualization, or planet-sized landscapes explorers (like Google Earth). The management (i.e. creation, editing, storage, transfer, processing and rendering) of large amounts of 3D data is a serious issue in graphics applications. We propose a new workflow, where the generation of scenes and textures is tightly coupled with rendering. PhD thesis of Charline Grenier.

ANR HDWorlds Huge Digital Worlds. Producing massive 3D models representing large scale virtual worlds with a high level of details is a major challenge in computer graphics. In industry as well, there is a strong demand for providing efficient algorithms to reduce manual authoring tasks. Procedural modeling and texturing (PMT) is known to provide a good solution to the scalability problem: it has excellent compression properties, it can produce large amounts of data with low user efforts and, by using stochastic processes, it can produce almost infinite varieties of data, based on a reduced set of parameters. In spite of all of these advantages, PMT still does not offer a suitable alternative to manual modeling, mainly because it is difficult to control, and because ensuring realism at various scales is a hard task. The goal of the HDWorlds research project is to overcome these limitations