Individual Colour Vision based Image Optimisation

'Individual Colour Vision-based Image Optimisation' (ICVIO) is a multidisciplinary project, which provides a step change in the ability to recover and predict quantities of individual human colour perception. It enables perceptually meaningful colour image optimisation and prediction of perceptual colour content in images between individuals with different colour vision, including normal, anomalous or deficient. Image optimisation is performed on digital images. The objective is to control colour fidelity and compliance, obtain pleasing imagery or compensate for changes in viewing conditions and observer etc. Today's methods rely on image independence, standard observers and viewing conditions. Thus, image rendition methods on given images make individual observers look at the same image. But as the visual systems are different between individual observers, they don't necessarily perceive the same image. We propose on-the-fly individualised image processing, such that the observers won't look at the same image, but rather perceive the same image. In ICVIO, the key is to link individual colour response (retinal sensitivity) with its according colour perception. The sensitivity is estimated with a few in situ colour discrimination experiments. The link is a unique transformation based on a novel colour metric (geometry), where perceptual discrimination and colour perception is characterised with consistency and accuracy. The metric is coherently intrinsic to the estimation of the sensitivity. The novel image processing technology in ICVIO applies scarcely explored classical science of human vision. It bridges human vision and digital imaging by providing imaging methods with significant potential for further scientific research and industrial application. The scope for benefit of ICVIO is broad: Art, heritage, medical imaging, robotics, surveillance, computer graphics, augmented and immersive virtual reality etc.

The project 'Individual Colour Vision based Image Optimisation' (ICVIO) aims at providing a step change in the ability to recover and predict quantities of individual human colour perception. The goal is to provide digital optimisation methods that are uniquely linked to individual colour perception of colour images. This will enable perceptually meaningful colour image optimisation and facilitate prediction of perceptual colour content in images between individuals with different colour vision. The key is to uniquely link individual cone fundamentals of human vision with a colour metric, in which said individuals' perceptual discrimination thresholds and fundamental colour perceptions can be characterised with high accuracy. We will propose novel methods of recovering individual cone fundamentals from a minimum of in situ colour matching and discrimination experiments. The colour metric will be based on a mathematical model whose geometry comprehensively describes the perceptual dimensions and thresholds for individuals, at a level which is beyond the capability of state-of-the-art methods. The individualisation of image optimisation will generate a need to advance existing image manipulation methods along with the opportunity to create new. It will encompass deficient, anomalous and normal colour vision. This multidisciplinary project links human vision, computational colour science, computer vision and digital imaging and provides imaging methods with a significant potential of industrial application.