提出了一种基于表面属性的三维模型检索方法——颜色分布。首先对颜色空间进行量化处理,然后对模型表面进行按面积均匀采样,最后统计采样点获取模型表面不同颜色的分布信息,从而形成颜色分布特征。该方法对模型的旋转、缩放以及噪声具有较强的鲁棒性。为验证颜色分布特征的检索效果,建立了一个包含1550个颜色模型的测试库。对测试库在多种相似性度量方法下进行颜色分布检索实验的结果表明:该方法对于基于表面属性的三维模型检索具有较好的检索效果。
In this paper we propose a surface property based content descriptor called color distribution. The descriptor is obtained from the statistic of the color values sampled from the models in quantized space. It is robust against scaling, rotation and noise. A test database including 1 550 colored models was created for evaluation. The proposed approach has been evaluated on the given database with several similarity measurement methods. Results show that the approach can give good results in terms of retrieval effectiveness for 3D model retrieval based on surface properties.