Statistical modeling of colour data

In this paper we investigate how best to model naturally arising distributi ons of colour camera data. It has become standard to model single mode dist ributions of colour data by ignoring the intensity component and constructi ng a Gaussian model of the chromaticity. This approach is appealing, becaus e the intensity of data can change arbitrarily due to shadowing and shading , whereas the chromaticity is more robust to these effects. However, it is unclear how best to construct such a model, since there are many domains in which the chromaticity can be represented. Furthermore, the applicability of this kind of model is questionable in all but the most basic lighting en vironments. We begin with a review of the reflection processes that give rise to distri butions of colour data. Several candidate models are then presented; some a re from the existing literature and some are novel. Properties of the diffe rent models are compared analytically and the models are empirically compar ed within a region tracking application over two separate sets of data. Res ults show that chromaticity based models perform well in constrained enviro nments where the physical model upon which they are based applies. It is fu rther found that models based on spherical representations of the chromatic ity data provide better performance than those based on more common planar representations, such as the chromaticity plane or the normalised colour sp ace. In less constrained environments, however, such as daylight, chromatic ity based models do not perform well, because of the effects of additional illumination components, which violate the physical model upon which they a re based.