Compression of color data in 3D semi-regular meshes

Abstract

In this paper, we propose a compression algorithm for color data in 3D semi-regular meshes. To enable the zerotree coding, an irregular mesh is preconverted to a semi-regular mesh via simplification and refinement. In the simplification, we use the color information to select collapsing edges adaptively, so that the color distortions of simplified meshes are alleviated. Moreover, in the refinement, we employ two color prediction methods: the butterfly prediction and the nearest prediction. It is shown that the nearest prediction provides more accurate prediction results. Simulation results on real 3D objects demonstrate that the proposed algorithm provides a promising compression performance.