Discriminative color descriptors

Color description is a challenging task because of large variations in RGB values which occur due to scene acciden-tal events, such as shadows, shading, specularities, illumi-nant color changes, and changes in viewing geometry. Tra-ditionally, this challenge has been addressed by capturing the variations in physics-based models, and deriving invari-ants for the undesired variations. The drawback of this ap-proach is that sets of distinguishable colors in the original color space are mapped to the same value in the photomet-ric invariant space. This results in a drop of discriminative power of the color description. In this paper we take an information theoretic approach to color description. We cluster color values together based on their discriminative power in a classification problem. The clustering has the explicit objective to minimize the drop of mutual information of the final representation. We show that such a color description automatically learns a certain degree of photometric invariance. We also show that a universal color representation, which is based on other data sets than the one at hand, can obtain competing performance. Experiments show that the proposed descrip-tor outperforms existing photometric invariants. Further-more, we show that combined with shape description these color descriptors obtain excellent results on four challeng