高级检索

几何曲面的双边滤波多尺度表示及几何细节增强

Multi-scale Surface Representation and Detail Enhancement Based on Bilateral Filter

  • 摘要: 为了有效地对几何模型进行编辑建模及特征检测等处理, 提出一种利用双边滤波对曲面进行多尺度表示及几何细节增强的算法.首先利用几何双边滤波对几何模型进行多尺度的分解, 获取其多尺度的表示和多尺度的细节层;然后对不同尺度的几何细节进行自适应的增强, 将这些增强的细节层与基曲面层进行重组, 重建出一个几何细节增强的几何模型.利用几何模型的多尺度细节抽取, 还提出一种几何细节迁移的算法, 将抽取出的多尺度几何细节分解映射到其他几何模型的基曲面上, 获取几何细节迁移的结果.文中算法实现简单、效率高, 能有效地增强几何模型的几何细节, 且整个几何模型不发生变形情况、保持几何特征.最后通过多个实例验证了该算法的有效性.

     

    Abstract: For efficient geometry editing, modeling, and feature detection, this paper presents a novel approach for multi-scale surface representation and detail enhancement in terms of bilateral filter.By the approach, we first decompose the geometry model using multi-scale bilateral filter, and extract the multi-scale detail layers.Then secondly, we reconstruct the geometry model with enhanced geometry detail by a combination of the enhanced detail layers and base surface.Using the extracted geometry detail, we propose a geometry detail transfer method by mapping the geometry detail of the surface model to the base surface of the target model.By our method, the geometry detail is enhanced while the geometry distortion can be prevented, with the geometry features well-preserved.The method is efficient and convenient in its implementation.The testing result demonstrates the effectiveness of the approach on a variety of geometry models.

     

/

返回文章
返回