Research And Application On Sharp Sensor Of Color Adaptation Model

With the revolution of imaging technology and widespread use of kinds of color media devices, in order to meet the needs of color transfer between different media and applications in industrial as digital video, now color matching of different media under different viewing conditions and environment has become a serious problem to be solved. Traditional colorimetric color description can only fit specific viewing conditions, apparently unable to cope with the new challenges. Color appearance model is a mathematical transformation model to simulate complex visual response of the human eye, and can be used to solve measurement problems of object color appearance in complex environments, it has become a hot topic of modern colorimety. Chromatic adaptation model is the core to build a color appearance model and can be separately applied to color reproduction and image applications and other fields.Sharp spectrally cone response is a research emphasis of modern color adaptation model, and it is basis of many modern chromatic adaptation models, it also known as sharp transform. Their common feature is the cone transformation does not happen in the physical cone space, but a narrower sharpened cone space, and the sensor corresponding to the cone receptors but distinguish from physiological is called sharp sensor. Finlayson et al. have obtained better performed and more sharpened senor named MSharp used spectral sharpening algorithm based on Lam & Rigg corresponding color dataset.Based on the same spectral sharpening algorithm put forward by Finlayson et al. we tried to use different and more color corresponding datasets in order to obtain higher precision sharp sensors, which may improve the overall prediction accuracy of color adaptation model. Three derivation programs then put forward, including different single datasets, combination datasets and reverse datasets, finally 34 sharp sensors obtained in total. Two highest precision sharp sensors named Tw’combin1 Tw’combin2 respectively derived from the two combination reverse datasets are picked out through a comprehensive evaluation, they have rather well performance when compared to other published models which are widely recognized and also do well in actual application evaluation tests. This is the main results of this research, in addition, the idea of mixed datasets and reverse datasets is the main innovation in this paper and is worthwhile of taking into further research, meantime provides a new train of thought for future research related to color adaptation.