Evaluating color texture descriptors under large variations of controlled lighting conditions

摘要

The recognition of color texture under varying lighting conditions is stillan open issue. Several features have been proposed for this purpose, rangingfrom traditional statistical descriptors to features extracted with neuralnetworks. Still, it is not completely clear under what circumstances a featureperforms better than the others. In this paper we report an extensivecomparison of old and new texture features, with and without a colornormalization step, with a particular focus on how they are affected by smalland large variation in the lighting conditions. The evaluation is performed ona new texture database including 68 samples of raw food acquired under 46conditions that present single and combined variations of light color,direction and intensity. The database allows to systematically investigate therobustness of texture descriptors across a large range of variations of imagingconditions.