The goal of this research line is to work on the definition of computational colour and texture representations. We focus on the basis of colour and texture perception, in order to automatize the task of giving judgements about the colours and the textures of an image. It can be useful for different computer vision applications involving colour and texture, such as, automatic image annotation, colour assessment on textured surfaces, estimation of illuminant, etc. To this goal, we work on four different research areas:
Colour constancy where we work on building algorithms to estimate scene illuminant from the image content.
Colour-Texture perception we work on computational models that simulates the chromatic induction of the human visual system.
Colour sharpening to get enhanced versions of colour images.
Colour Naming to simulate the task of assigning a colour name or category to an specific image segment.
Texture description where we work on blob detection, as the basis for a high level and non-numeric description of textures.
Colour & Texture grouping to define mechanisms for segmenting images according to the chromatic and spatial relationships of the image elements.