Composite Color Invariant Feature H′ Applied to Image Matching
説明
Object color is one of the most important feature in camera-based object matching. Color invariants are features based on the models of color observation that tends to be constant under varying conditions of illumination and surface. In this work, we analyze the estimation process of the color invariants by Geusebroek et al. from RGB images, and propose a novel invariant feature H′ based on the elementary invariants to meet the circular continuity residing in the mapping between colors and their invariants. The use of the proposed invariant in combination with luminance, contributes to improve the retrieval performances of partial object image matching under varying illumination conditions.