Influence of Background in the Colour Appearance of Images
Résumé
In this paper we have investigated whether the colour perception is affected by the distribution and the spatial organization of colours in a complex image. In the first part of our study, we analysed the influence of scene content and of background types on the colour appearance judgement. To reach this aim we ran visual assessment experiments based on the magnitude estimation technique and we investigated various visual phenomena, such as brightness adaptation, chromatic spatial adaptation, contrast effects due to sizes and colored backgrounds, display field sizes and dynamic range in the scene. The examination of numerous visual assessment results done showed that the influence of the background on colour appearance is more noticeable for complex images with high frequencies than for color images with low frequencies or simple images. Likewise, the influence of the background on colour appearance is more noticeable for chromatic images with a large gamut than for less colored images with a low gamut. In the second part of our study, we analysed the influence of the local colour saliency on the colour appearance. To reach this aim we developed a computational model to measure colour contrast. Our motivation was to define an objective metric consistent with observer valuation. The proposed model integrates in a single model the influence of average colour perception and the interactions between local and global spatial structures according to the visual eccentricity. The measure of colour contrast relies on a set of parameters organized in a hierarchical structure. The computation is based on spatial criteria and integrates low-level factors calculated on defined regions relatively to their local and global neighbourhoods.