Old Positions and Internships

This is the archive page. Back to the current offers page.

INDEX: Types of positions and internships

POST-DOC

Artefactology: Characterization, perception and minimization of visual artifacts in paradoxical advected-preserved textures

see URL....   »More information

Lazy multiscale matter distribution and lighting for realistic walkthrough the galaxy in real-time

see URL....   »More information

Efficient and realistic rendering of image-based models

Our goal is to improve the realism of Image Synthesis, by including real-world, complex objects. Through a cooperative project, we have access to high-quality image-based models. In this post-doc, we will focus on rendering them, inside classical geometrical scenes.

Our work will focus on efficient and realistic rendering of image-based models, including lighting, shadowing and re-lighting effects. We will focus our research in three major directions:

  1. efficient rendering of image-based models, with a rendering cost proportional to the visual complexity of the model, not to the image resolution;
  2. self-shadowing and masking effects: just like parts of the models are hiding each other in the acquisition process, they are also hiding each other in the rendering process. This has an influence both on the picture of the model (masking effects) and on the illumination of the model (self-shadowing effects);
  3. re-lighting of the models: the image-based model was acquired under specific lighting conditions. For a realistic rendering under different lighting conditions, we need to model the influence of lighting on the image-based model, including both direct and indirect lighting.

Postdoc: MIRO: Methods for the legible and Interactive Rendering Of 3D complex data

In this project, we are interested in using NPR for the interactive and legible rendering of complex 3D scenes. Many examples in scientific visualization prove that photorealism does not always offer meaningful images. For example in medical books, the representation of organs is more frequently based on sketches and annotations than on photographs. The postdoc will on these topics, in the context of archeology, museography, medical simulation, all applications that require specialized visualizations. This postdoc will take place 12 months in Grenoble (ARTIS team) and 6 months en Bordeaux (IPARLA team) and is funded by INRIA. ...   »More information

PhD Thesis

Dimensional Analysis of Light Transport

This PhD aims at exploring light transport through the transversal view of light transport operators between the infinite dimensional function spaces of light fields (imagine, e.g., reflectance as an operator that transforms a distribution of incident light into a distribution of reflected light). In addition to be linear in these spaces, these operators are also very sparse. As a side effect, the sub-spaces of light distributions that are actually relevant during the computation of a solution always boil down to a low dimensional manifold embedded in the full space of light distributions. Reflectance over ''smooth'' materials for instance, converts incident illumination into a low dimensional set of reflected light distributions. Rather than using linearity to compress precomputed radiance transfer (PRT), we want to leverage recent ad-hoc dimensionality reduction methods: Compressive sensing, Gaussian processes, Monte-Carlo eigenanalysis and unsupervised machine learning from huge volume of generated data. Expected results are two-fold: (1) a theoretical understanding of the dimensionality and structure of light transport operators (examples include scattering, reflectance, global transport in glossy scenes). Possible applications will cover re-lighting, material acquisition and efficient simulation of complex light transport problems. (2) We also envision that controlled approximations of the low dimensional operators will open the way to efficient approximations of light transport which may also be suitable for real time applications such as interactive lighting for video games. The PhD candidate should have a solid background in mathematics and programmation. Knowledge in GPU programming is a plus. ...   »More information

Dimensional Analysis of Light Transport Operators

Multiscale real-time procedural animation and rendering of oceans

see URL...   »More information

Multiscale real-time procedural animation of clouds and cloud systems

see URL...   »More information

Acquisition de propriétés de matériaux dans un contexte de réalité augmentée

Le processus classique en synthèse d’images consiste à créer un modèle géométrique des objets, puis à y a jouter des propriétés de matériaux, et enfin à calculer la simulation de l’éclairage sur ce modèle. Ce processus limite le réalisme, notamment en ce qui concerne les modèles issus du monde réel.

Notre objectif est de permettre l’acquisition des propriétés de réflectance d’objets issus du monde réel, afin de permettre leur placement dans des mondes virtuels. Les applications concernent aussi bien les effets spéciaux dans le cinéma que le réalisme dans les jeux vidéos.

Cette thèse, financée par la région Rhône-Alpes, est réalisée en collaboration avec le laboratoire LIRIS, et le doctorant sera à 50 % de son temps au LIRIS.

Rendu Temps-Réel de Modèle Complexes

Nous nous plaçon dans un projet coopératif, où nous avons accès à des modèles complexes, dont la géométrie est acquise par exemple par scanners 3D.

Ces modèles sont trop complexes pour pouvoir être utilisés dans des applications de rendu temps-réel. Il est donc nécessaire de les simplifier et d'inventer de nouvelles représentations, dites représentations alternatives, qui permettront une représentation en temps-réel tout en conservant les détails géométriques fins que donne l'acquisition.

Dans une deuxième étape de travail, nous voulons simuler l'éclairage sur ces représentations alternatives, en commençant par l'éclairage direct et les ombres, puis l'éclairage indirect.

Notre objectif est de pouvoir inclure ces objets dans des scène en Image de Synthèse, en conservant leur aspect réaliste.

Simulation de l'éclairage en temps-réel pour jeux vidéos

Le projet GENAC (Génération Automatique de Contenus) porte sur le développement d'outils pour le jeu vidéo : animation et rendu temps-réel. Ce projet, porté par les équipes ARTIS et EVASION de l'INRIA Rhône-Alpes, en commun avec les industriels Eden Games et Widescreen Games et le LIRIS, a été labellisé par le pôle de compétitivité Imaginove, et est financé par le Fonds de Compétitivité des Entreprises, la Région Rhône-Alpes et le Grand Lyon.

Les algorithmes de recherche développés dans le cadre de ce projet seront intégrés à une plate-forme de jeux vidéos fournie par Eden Games.

Dans le cadre de ce projet, l'INRIA Rhône-Alpes recrute un Doctorant, qui travaillera dans l'équipe ARTIS, sur la simulation en temps-réel des effets d'éclairage global.

But de la thèse : Développer un algorithme d'éclairage global en temps-réel, capable de répondre aux contraintes du jeu vidéo : affichage en temps-réel, scènes de grande taille, structure de scène imposée par le moteur de jeu, scènes animées. Cet algorithme sera intégré dans le moteur de jeu vidéo fourni par les partenaires industriels, le travail d'intégration étant effectué, en partie, par l'ingénieur recruté également dans le cadre du projet. Le travail est décomposé en sous-tâches : simulation de l'éclairage indirect reçu par un object animé au sein d'un environnement statique, placement automatique des points d'échantillonnage, simulation des variations de l'éclairage liées aux saisons.

Modelling Complex Light Sources for Image Synthesis

PhD position at GRAVIR/IMAG - Grenoble - France
Context: Image Synthesis and PhotoRealistic Rendering
In this thesis, we want to model complex light sources, such as fire, car headlights or a cristal chandelier... We envision a situation where the exact geometry of the light sources is either unknown (as in the case of fire) or too complex for a complete numerical simulation (as in the case of the cristal chandelier). We want to measure the lighting from these light sources and to store them in a compact model that expresses all the effects of the light source, either at close range or for long distances. For light sources with a spatial extension, illumination at close range will have a spatial dependancy.
We attach a great importance to the spectral content of the lighting, which we also want to store. Most light sources will also have a temporal variability. The models obtained through this study and modelling will be used in real-time rendering, global illumination simulations and virtual reality applications.
Candidates must be eligible for being granted a Marie Curie fellowship. In the following,European citizens include all people from European member states and from associated states. - See Eligibility Criteria -.

...   »More information

INGENIEUR

Interactive procedural Gabor noise on surfaces revisited

see URL...   »More information

Moteur de rendu Gigavoxels

CDD Ingénieur. See URL:...   »More information

Moteur de rendu Gigavoxels

CDD Ingénieur. See URL:...   »More information

Master2 Rech (M2R, ex DEA)

Smart brushes for 3D digital painting

...   »More information

Sculpting procedural noise

See URL: ...   »More information

Playing with Fourier spectrum for real-time ocean explorer and texture generation

see URL:...   »More information

Real-time walk-through the Milky-Way: continuum to points, on-the-fly generation, and mixing with catalog

see URL:...   »More information

Shading and shadowing in panorama maps

...   »More information

Analyzing correlations between shape/light/materials

...   »More information

2D element texture design

...   »More information

Real-time global illumination using a Latent Variable Model

The goal of the internship is to experiment with the interpolation of global illumination using a Gaussian Process latent variable model. The latent variables will be chosen so as to optimize the quality of the interpolation. Global illumination usually requires a costly calculation (as opposed to direct illumination). Based on the intuition that the two are linked by a low dimensional operator (despite the infinite dimension of the space direct and global illumination potentially belong to), we want to find low dimensional approximations of this relationship. Since linear parameterizations (using e.g. PCA) have some limits (in particular because they cause visual artifacts) we aim for non linear parameterizations. This naturally leads to looking into Gaussian processes. The student will essentially need to * get familiar with the mathematical principles of global illumination and Gaussian processes; * get familiar with the existing Maverick’s rendering platform; * implement the precomputation of images as well as the Gaussian process training in image space; * implement the real-time rendering loop in image space; * create a nice proof-of-concept demo in a non trivial geometric scene; * (optional) implement an object-space–videogame compatible–version where the light is stored into textures * (optional) discuss and experiments the advantages of different function bases to generate coefficients for direct illumination; The internship requires to be fluent in C++ and very comfortable with the acquisition of new mathematical concepts. Do not hesitate to come talk with me. If you have your own ideas about global illumination and what can be done, I'm open as well. mail: cyril.soler@inria.fr....   »More information

Real-time global illumination using a Latent Variable Model

See url....   »More information

Playing with Fourier spectrum for real-time ocean explorer and texture generation

See URL:...   »More information

Procedural modeling of 3D realistic galactic dust and nebulas

See URL: ...   »More information

Real-time walk-through the Milky-Way: continuum to points, on-the-fly generation, and mixing with catalog

See URL:...   »More information

Revisiting volumetric ray-tracing to make it well-posed

See URL:...   »More information

2.5D Procedural Stylization

...   »More information

Inverse Procedural Textures

Procedural texture are images generated by a tree of mathematical operations/objects such as random noise, clamping, masking, normal maps, shading, etc. They are used to create stationary textures that show some structure. Existing tools such as substance (Allegorithmics, https://www.allegorithmic.com/) to design procedural textures allow a huge variety of results. The goal of this internship is to create an algorithm for recovering a procedural graph that can generate an image that is as close a possible to an input picture. This is a very difficult problem, which we will segment into simpler tasks, and leverage the recent advances in deep learning, for image comparison and image classification....   »More information

Parameterization of the Indirect Illumination Manifold

The simulation of global illumination consists in numerically solving an integral equation that expresses the equilibrium of light energy in a scene. For practical reasons, one usually separates the direct illumination, that denotes light directly comes from the light sources and bounces on a single object before reaching the camera, from the indirect illumination, which may bounces an arbitrary number of times in the scene before reaching the camera. The distribution of indirect illumination depends very smoothly on the positions of the light sources, and therefore only occupies a restricted sub-space of the full span of possible distributions of light energy in a scene, when illumination conditions are changed. Consequently, we believe that it is possible to compute a low dimensional parameterization of the manifold of indirect illumination distributions. We would like to use this parameterization to perform real-time calculations of global illumination in video games that react to the players position and actions, as well as performing relighting tasks into static images. ...   »More information

Exploring Difficult Manifolds for Light Transport

To produce photorealistic pictures, we seek all paths connecting light source and camera, after multiple bounces on the scene geometry.
Explicitly finding all paths is difficult, especially with multiple refractive interfaces. Manifold Exploration is the idea that the set of good paths forms a manifold in path space, and once we have a point on this manifold, we can explore path space while staying on it.
Manifold exploration is great for exploring difficult paths, but we have to find a point on the manifold first. With two specular interfaces (especially refractive), this becomes difficult. Our goal in this project is to extend manifold exploration to these difficult cases. ...   »More information

Adapting the Combined Diffraction-MicroFacet Model for designers

The realism of photorealistic pictures depends a lot on the realism of the materials used to depict the scene. The micro-facet model is widely used in research and industry for its simplicity and versatility. It covers materials from diffuse to specular, and connects to physical parameters (material chemical composition, roughness). In Summer 2017, we introduced a new model that extends the micro-facet model by taking into account diffraction effects in the microgeometry of the surface. The new model is better at representing the observed behaviour of actual materials. While the new model is good at representing actual materials, it was not designed for artists and designers. The goal in this research project is to adapt it for practical users: professional artists and designers: simplification, intuitive parameters, filtering......   »More information

Ageing and Reverse-Ageing for Cultural Artifacts

The realism of photorealistic pictures depends a lot on the materials used to depict the scene. When we depict actual objects, it is good to have a connection with the physical properties: chemical composition, surface roughness, etc. It is especially interesting when we deal with cultural artifacts, such as paintings, sculptures, metallic artifacts. What we have is how they look now, and their current chemical composition. What we would like to have is how they looked when they were new (reverse-ageing). Some ageing effects are related to oxidation; the effects can be just a change in the chemical properties of the material, like tin transforming into tin oxide. In more extreme cases, there is also a change in the surface geometry, like iron transforming into rust. Our goal in this project is to take models of ancient cultural artefacts, such as tin vases or bronze swords, and reverse the effects of time to produce photorealistic pictures of the artefacts as they were when they were new. ...   »More information

Rendering​ ​ panorama​ ​ maps​ ​ in​ ​ the​ ​ “atelier​ ​ Novat”​ ​ style

The Atelier Novat was born in the '60s, founded by Pierre Novat and became the French standard for mountain panorama: http://atelier.novat.free.fr. Our goal in this project is to reproduce the atelier Novat style using digital means. ...   »More information

Connecting Chemical / Physical Properties and Material Appearance

The aspect of a material depends on its surface characteristics and on its chemical composition. The dependency on surface characteristics, such as surface roughness, is relatively well known. The connection with the chemical composition is less well known. When a material is reduced in a powder, its color changes due to Mie scattering. Overall color depends on grain size as well as chemical composition. When the powder in included in a solvant (for oil painting), color also depends on solvant properties and concentration. Physical models explain the connection between grain size, concentration and color. The model is (obviously) non-linear. It is also computationally expensive. The goal in this project is to predict the visual properties of a mixture of hematite powder and solvant, based on the physical properties of the mixture (grain size, chemical composition, solvant). This project will be done in cooperation with Institut Néel, allowing access to the actual properties of the materials, along with its color....   »More information

Example-Based Texture Transfer Between Images

See URL....   »More information

Material Edition in Images

See URL....   »More information

Procedural modeling of multiscale galactic dust

see URL: ...   »More information

Interactivel Change of Materials in Images

Our goal is to build a tool to consistently replace materials on objects in static images, where the term material denotes the properties of surfaces that define how the surface interacts with light (Typical use case: take a picture of a car, replace the car paint). The challenge in this project is to be able to extract a consistent pair of geometric properties and incident illumination in a single image from the unique knowledge of the material properties of objects in the images. This type of problem generally falls into the shape from shading category, which is known to be ambiguous. However, in our situation, we do not seek the real geometry of the objects nor the real distribution of incident light but only a consistent pair that would allow to replace the material properties while keeping the same viewing and illumination conditions. The project has already started, and we have a working pipeline for convex objects. In real life cases however, objects are not convex and require to also compute which parts of incident half-sphere of directions each pixel actually sees. We propose to compute this visibility information, in order to increase the accuracy of the extracted information. ...   »More information

A new procedural primitive for fractal galactic dust

see URL...   »More information

Studying and solving the strange visual artifacts occuring when procedural texturing with paradoxical requirements

see URL...   »More information

Top-down texture painting

see URL...   »More information

Example-Based Material Transfer Between Images

see URL....   »More information

Photographic Lighting Design

...   »More information

Modeling the emerging effects of sub-voxel data for scalable realistic realtime rendering of GigaVoxels scenes

See here:...   »More information

Procedural generation of details for the realistic real-time rendering of galaxies

See here:...   »More information

Global appearance of non-uniform volumes of stars and dust clouds for the realistic real-time rendering of galaxies

See here:...   »More information

Exaggerating using vector fields: tools and applications

voir URL / see URL...   »More information

Design of flows on smooth surfaces

...   »More information

Dimensional Analysis of Light Transport Operators

Real-time Animation and Rendering of Ocean Whitecaps

...   »More information

Understanding the temporal coherence problem for the stylization of animations

...   »More information

Temporally coherent stylization of 2D animations

...   »More information

Simulating an evolving cloud sky

...   »More information

Amplifying smoke simulation with animated details

...   »More information

Mesh Segmentation by Diffusion

The idea is to leverage from an existing image segmentation technique called Diffusion Segmentation, to be able to segment meshes....   »More information

A voxel engine to render ultra-complex animated scenes in real-time

Engineering project....   »More information

A journey in a procedural volume / Voyage dans un volume procédural

...   »More information

Simulating an evolving cloudy sky / Simuler un ciel nuageux évolutif

...   »More information

Textures 3D pour l'animation aquarelle

Un des chalenges du Rendu Non Photorealsite (NPR) est de reproduire l'apparence de mediums traditionnels comme l'aquarelle, le fusain, la peinture... Une des méthodes employées consiste a reproduire certaines des caractéristiques du medium avec des textures : texture de pigments et papier pour l'aquarelle, texture de coups de pinceaux pour la peinture...

Dans le cadre d'une animation NPR, la difficulté consiste a faire bouger ces textures en accord avec les mouvements de l'animation, tout en conservant l'aspect plat et homogène du medium traditionnel. Le sujet de ce stage est d'étendre la méthode de « dynamic canvas » [2] a l'animation d'objets 3D. Dynamic canvas permet d'animer une texture de papier 2D selon les mouvements d'une camera. Le but du stage serait d'appliquer un mécanisme similaire a des textures volumiques selon le mouvements d'objets 3D. Pour cela une des pistes de recherche serait d'utiliser les recents travaux en synthese de texture [3] pour generer des textures volumique (papier, pigment, coup de pinceaux) qui respectent les contraintes de Dynamic Canvas.

References

[1] http://artis.imag.fr/Publications/2007/BNTS07/

[2] http://artis.imag.fr/Publications/2003/CTPDGD03/

[3] http://www.johanneskopf.de/publications/solid/index.php

...   »More information

Paramétrisation 3D/2D avec cohérence temporelle de lignes pour rendu expressif

Dans un rendu expressif construit de lignes à partir d'une scène 3D, les effets sont paramétrés le longs de ces lignes. Mais lors d'une animation, il est difficile de garder une paramétrisation cohérente à cause des déformations 3D des lignes mais aussi des occlusions. Le sujet de ce stage est de définir comment paramétrer une lignes issue d'un modèle 3D (silhouette par exemple) pour lui affectée une stylisation. L'idée principale est que les différents type d'effet que l'on peut appliquer à une ligne ont une paramétrisation plus appropriée que d'autre (2D/3D/autre ...). Il faut aussi définir comment faire évoluer les paramétrisations au cours de l'animation de manière cohérente. Une fois ces paramétrisations identifées, l'étudiant devra mettre en place un suivit temporelle des différentes paramétrisations et dévolopper un moteur de rendu expressif de scnène 3D avec ces paramétrisations. ...   »More information

Optimisation du placement de caméras pour la reconstruction de personnages

Dans le cadre du projet CYBER-II, nous cherchons à trouver la meilleure position des cameras pour reconstruire au mieux la structure 3D de l'animateur...   »More information

Affichage d'objets lointains par grille de points

Le but de ce sujet est de mettre au point une représentation alternative adaptée à l'affichage d'objets très lointains, ne couvrant que quelques dizaines de pixels à l'écran. ...   »More information

Ombre douces temps réel

Dans le cadre du projet CYBER, nous cherchons à produire des ombres douces en temps réel en utilisant au mieux les capacités des cartes graphiques. ...   »More information

Dessin au trait en 3D

Définition d'un outils de dessin sur modèle 3D selon divers points de vue avec interpolation entre les points de vue....   »More information

Stylisation d'une animation à partir d'une séquence de G-buffers (plusieurs sujets, divers niveaux)

Nous voulons proposer un système permettant à un artiste de styliser une animation qu'il aurait faite avec un modeleur classique. Pour cela nous avons choisi une approche qui travaillerait avec la succession des images de l'animation augmentées de G-buffer (geometry-buffer [saito-Takahashi]) qui sont des images refletant des informations 3D de la scène : image de profondeur, image des normales, etc .......   »More information

Style Analogies

Application de la méthode des 'image analogies' a des informations de style de haut niveau, afin d'être capable de capturer et de reproduire la façon dont un modèle est rendu en NPR....   »More information

Visibilité avec Occludeurs Discrétisés

Le but est de précalculer la visibilité entre des régions de l'espace, par exemple pour pouvoir répondre à la question "Quelle partie d'une pièce sont éclairées directement par un lampe sphérique (et non ponctuelle)?". Le sujet fait appel à l'intuition géométrique et nécessitera autant de travail d'implémentation que de recherche "sur le papier"ou devant le tableau....   »More information

Ray-tracing de Height Field sur le GPU

On cherche à augmenter le relief des scènes 3D en ajoutant un champ de hauteur au textures et en utilisant le GPU pour rasteriser ce champ....   »More information

MAITRISE

webMobiNet : plateforme pédagogique de programmation de simus et jeux en réseau

cf URL....   »More information

Real-time optical effects in OpenGL using GigaVoxels

Master or Engineering see URL...   »More information

Collision Detection on GPU for real-time applications

Master or Engineering see URL...   »More information

Self-similar textures design tool

...   »More information

Advice for prospective international students

Internships for international students are processed through the program "Internship at INRIA for international students" http://www-direction.inria.fr/international/PROGRAMMES/internship/index.en.h tml As explained in the description of this program, we only accept application through the request of a local sponsor, usually a university faculty or dean, for which the current list is at http://www-direction.inria.fr/international/PHP/Internship/AffichIIeng.html Please feel free to contact the local sponsor at your university if you wish him or her to sponsor your application with us.

Simulation de Peinture Animée

La peinture animée est traditionnellement effectué en photographiant une plaque de verre sur laquelle l'artiste peint à l'acrylique (généralement avec les doigts) chacune des images constituant l'animation. Le sujet proposé est la mise en place d'une simulation de ce procédé. Le rendu sera guidé par un film en entrée, vidéo ou 3D (accompagné de renseignements complémentaires tel que le flux optique) et l'utilisateur contrôlera le style du rendu en définissant les action à accomplir pour peindre l'animation. Il s'agit donc de définir un ensemble d'action que l'on peut faire pour modifier une image en s'inspirant de ce qu'un artiste peut faire à la main : ajouter de la peinture, étaler la peinture, mélanger les couleurs. Et de trouver comment combiner ces actions pour représenté le film en entrée. Pour amélioré l'aspect final, il est important d'explorer les pré et post traitement que l'on peut effectuer sur le rendu. Par exemple faire une quantification/abstraction du film en entrée. ...   »More information

Advice for prospective international students

Internships for international students are processed through the program "Internship at INRIA for international students" http://www-direction.inria.fr/international/PROGRAMMES/internship/index.en.h tml As explained in the description of this program, we only accept application through the request of a local sponsor, usually a university faculty or dean, for which the current list is at http://www-direction.inria.fr/international/PHP/Internship/AffichIIeng.html Please feel free to contact the local sponsor at your university if you wish him or her to sponsor your application with us.

LICENCE

Advice for international students

Internships for international students are processed through the program Internship at INRIA for international students

As explained in the description of this program, we only accept application through the request of a local sponsor, usually a university faculty or dean. Here is the current list of partner institutions and sponsors

Please feel free to contact the local sponsor at your university if you wish him or her to sponsor your application with us.