27 juin 2016
Maison de la recherche, Paris
Registration required/inscription nécessaire
9:30-10:45 – Olivier Massin
11:00 – 11:45 – Alban Flachot
12:00-14:00 – Lunch Break
14:00-15:15 – Yasmina Jraissati
15:30-16:30– Christoph Witzel
16:30-17:45 – Igor Douven
Intervenants & résumés :
Igor Douven (SND)
Abstract : Regier, Kay, and Khetarpal  report the results of computer simulations that cluster color stimuli on the basis of their coordinates in CIELAB space, one of two commonly used perceptual color spaces. Regier and coauthors find partitions of those stimuli that are strikingly similar to the way actual color lexicons partition color space. They do not argue for the custom-made clustering method used in their simulations, nor for the assumption of CIELAB space. The present paper aims to answer the question to what extent their computational results depend on these assumptions. It does this by applying a great variety of known clustering methods to Regier et al.’s stimuli, and by assuming not only CIELAB space but also CIELUV space, the other main color space.
Alban Flachot (Paris-Descartes, Laboratoire de psychologie de la perception)
Color naming, unique hues and singularity in reflection properties : Discussion of a model and further developments.
Abstract : Cross cultural studies show that red, yellow, green and blue are color categories common to most languages throughout the world. Given that the prototypes of these categories also correspond to the so-called unique hues, this pattern in color naming suggests that these four colors and the surfaces they are associated with have a particular perceptual status.
A recent model by Philipona & O’Regan (2006) shows that the information about the reflected light from a surface that is accessible to the three human photoreceptors is linearly related to the information accessible about the illuminant. The characteristics of this linear relation, in particular its singularity in the mathematical sense, bears a remarkably high correlation with the tendency for the surface color to be given a name across many cultures. Singularities in the linear relation also accurately predicted unique hues and hue cancellation. Recent studies have contributed to developing the model further : Vazquez-Corral et al (2012) proposed revisions to make conceptual links with common approaches to color constancy by considering a spectrally sharpened sensor basis. Witzel et al (2015) further investigated how the singularities under consideration depend on the illuminant, the reflectance functions of the surfaces and the cone sensitivities.
This talk will offer a review and discussion of these recent studies. A new method to compute the action of a surface on incoming light will be proposed, giving a new insight into the validity and limits of the linear hypothesis. Alternatives to spectral sharpening as well as attempts to further develop the model will also be presented.
Yasmina Jraissati (American University of Beyrouth)
Color Categories in context
Olivier Massin (Université de Genève)
Colours as stuffs
A common assumption within colour philosophy and colours sciences is that colours belong to the ontological category of properties. I shall alternatively defend that phenomenal colours primarily belong to the category of stuffs.
Christoph Witzel (Giessen University, Department of Psychology)
Perceptual organisation of colour categories or categorical organisation of Munsell chips ?