点击次数：

论文类型：期刊论文

发表时间：2013-11-01

发表刊物：Journal of Information and Computational Science

收录刊物：EI、Scopus

卷号：10

期号：16

页面范围：5135-5143

ISSN号：15487741

摘要：Geometry image is an important representation for geometry models. High compressibility will be beneficial for preserving more graphics information when geometry images are compressed. In this paper, we propose a mesh parameterization method aiming to increasing the compressibility of geometry images that generated from open meshes. Firstly, we design a compressibility aware energy of the mesh by studying the relationship between the one-ring neighbors of the mesh vertex and the pixel neighbors of the image point. Secondly, we solve the parameterization problem by minimizing the energy defined as a weighted sum of a compressibility aware term and a conformal term. As a result, a geometry image with high image compressibility can be constructed by uniformly resampling the parameterized mesh in its parameter domain. Experimental results illustrate that the proposed method always achieves better local linear correlation and lower reconstruction error for different sampling resolutions of geometry images. ? 2013 Binary Information Press.