• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International conference on computer analysis of images and patterns;CAIP 2011 >A Robust Approach to Multi-feature Based Mesh Segmentation Using Adaptive Density Estimation
【24h】

A Robust Approach to Multi-feature Based Mesh Segmentation Using Adaptive Density Estimation

机译：基于自适应密度估计的基于多特征网格分割的鲁棒方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In this paper, a new and robust approach to mesh segmentation is presented. There are various algorithms which deliver satisfying results on clean 3D models. However, many reverse-engineering applications in computer vision such as 3D reconstruction produce extremely noisy or even incomplete data. The presented segmentation algorithm copes with this challenge by a robust semi-global clustering scheme and a cost-function that is based on a probabilistic model. Vision based reconstruction methods are able to generate colored meshes and it is shown, how the vertex color can be used as a supportive feature. A probabilistic framework allows the algorithm to be easily extended by other user defined features. The segmentation scheme is a local iterative optimization embedded in a hierarchical clustering technique. The presented method has been successfully tested on various real world examples.
机译：在本文中，提出了一种新的且鲁棒的网格分割方法。有多种算法可在干净的3D模型上提供令人满意的结果。但是，计算机视觉中的许多反向工程应用程序（例如3D重建）都​​会产生非常嘈杂的数据，甚至是不完整的数据。提出的分割算法通过鲁棒的半全局聚类方案和基于概率模型的成本函数来应对这一挑战。基于视觉的重建方法能够生成彩色网格，并显示了如何将顶点颜色用作辅助功能。概率框架允许该算法容易被其他用户定义的功能扩展。分割方案是嵌入层次聚类技术中的局部迭代优化。所提出的方法已在各种实际示例中成功进行了测试。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号