Abstract
In real images, color segmentation is a challenging problem because of natural variations in light and shadows. Most existing color models have limitations in distinctive representation of colors in real scenes. Color segmentation approaches are highly dependent on the robustness of the used color model to these variations. In this paper, we present a novel color representation for color segmentation. The proposed model is a fusion of classical color representations. It is calculated using a closed form linear transformation from a high dimensional classical feature space to a lower dimensional new space. We prove the high performance of the proposed model in terms of robustness to natural variations and distinction between different colors, when compared with classical color models. Through color segmentation of real images, we provide a quantitative and qualitative comparison between the proposed model and classical models. The experimental results support the potential of the proposed model.