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


( see video )

1-2 year Postdoctoral position
at Grenoble, France, at INRIA

contacts:  Fabrice NEYRET - mail: Fabrice.Neyret@imag.fr
Joelle THOLLOT – mail: Joelle.Thollot@imag.fr
Romain VERGNE – mail : Romain.Vergne@inria.fr

applying to this position : see HERE

Subject

In several domains in the field of Computer Graphics and Visualization, one encounters the same paradoxical requirement of having a pattern that has to simultaneously keep constant aspect characteristics and accurately follow a motion field.  The pattern might be procedural noise texture [ADV], sampling of fluid particles [fluids], or  paint marks and canvas in expressive rendering [NPR]. The characteristics to preserve might be scale, spectrum, histogram, momentums, or windowed combinations. The motion field might be screen-space apparent optical flow of a 3D scene, or fluid motion.
Moreover, temporal coherency is required: besides appearance preservation and motion conformance, one wants to avoid all and any perceivable visual artifact either discontinuous (popping) or continuous (clandestine secondary motion). Many dedicated models have been proposed in each field, but the induced secondary motion is the fundamental and difficult issue (almost unstudied), since it deals with what our visual system perceive or not of the emerging unwanted consequences of paradoxical main requirements.

Clandestine secondary motion is already obvious in very basic cases such as infinite zoom in black and white Perlin noise (cf image), i.e., thresholded Perlin noise where the components advected by the 2D radial motion (zoom) are progressively replaced so as to preserve the noise statistics. This allows to settle a convenient experimental bench to test the interactions between various algorithmic combinations, textural characteristics, and families of perceivable artifacts. Starting with characterizing and organizing the families of artefacts, to somehow initiates a domain of artefactology :-) .

A part of the work will consist in experiments with the bench above, testing the perceptual effect of various ranges and kinds of statistic preservation or violation together with objective measures, bibliography in perception studies so as to make educated guesses, possibly some cheap perception experiments, but also continuous care about the right control and meaningfullness of the bench, tested parameters and criterions.

Beside the beginning of a general understanding of the issue, we expect from this studies to get recommendations helping to improve time coherency in various applicative algorithms: Which kind of transformations are unperceived or extremely perceived, or which range of parameters make them so.  Plus possibly, slight key-changes in tools, definitions, requirements, normalizations,  so as to avoid ill-posed problems.

References:
[ADV] Advected textures: http://hal.inria.fr/inria-00536064, http://www-evasion.imag.fr/Publications/2003/Ney03/
[fluids] Scalable river animation:
http://maverick.inria.fr/Publications/2009/YNBH09/
[NPR1] Animated paints:
http://hal.inria.fr/inria-00510250/ , http://hal.inria.fr/inria-00463592/
[NPR2] Animated stylisation: 
http://hal.inria.fr/inria-00487986/ , http://hal.inria.fr/inria-00636210/

Profile



Applying to this position : see HERE