Independence of luminance and chrominance information in natural scenes strongly depends on data pre-processing. - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2016

Independence of luminance and chrominance information in natural scenes strongly depends on data pre-processing.

Résumé

Color vision is an impressive ability of human visual system. Three different types of cones in the retina, and then three postreceptoral cone-opponent mechanisms enable us to extract both luminance and chromatic information from our environment. Theories of efficient coding argued that the visual system evolved this way to optimally process natural scenes. Therefore, the study of statistical properties of this class of images has great potential value to expand our understanding of visual perception. Several studies investigated luminance and color joint distribution in images and claimed that luminance and chromatic edges were not independent of each other (Fine et al. 2003) and that most edges were defined by luminance contrast with color information being redundant (Zhou & Mel 2008). In 2009, Hansen and Gegenfurtner showed using mutual information that luminance and chromatic edges constitute independent sources of information and that their independence increases along successive stages of visual processing. Here we replicated their analysis in over 1000 images taken from the McGill calibrated color image database (Olmos & Kingdom, 2004) and found that the results concerning mutual information strongly depend on several parameters, such as cone fundamentals choice, luminance definition or data pre-processing. It suggests that separating luminance and chrominance information in natural images is not a trivial task and may have important implications on results concerning their dependency and their respective functions in visual processing.
Fichier non déposé

Dates et versions

hal-01836558 , version 1 (19-07-2018)

Identifiants

  • HAL Id : hal-01836558 , version 1

Citer

Camille Breuil, Simon Barthelme, Nathalie Guyader. Independence of luminance and chrominance information in natural scenes strongly depends on data pre-processing.. Groupe de Recherche Vision Annual Meeting, Nov 2016, Toulouse, France. ⟨hal-01836558⟩
126 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More