3-D Object Tracking Method Robust to Illumination Change

Abstract

In case of tracking an object using visual cues,illumination changes due to intensity of light sources,light directions, and shadows seriously affect trackingperformance. Especially, this problem is critical whendetecting and tracking a 3-D object because the RGBimage is formed from not only illumination andreflectance but also shape of the target object. Use ofchromaticity image can be is a solution of thisproblem since it reflects only intrinsic characteristicof subjects. In this paper, we propose a 3-D objecttracking method robust to illumination change basedon whole channels of chromaticity image. Theproposed method shows robust tracking performancedespite illumination change and specular reflection.