...
首页> 外文期刊>電子情報通信学会技術研究報告. パターン認識·メディア理解. Pattern Recognition and Media Understanding >Retrieval Method for 3D Object Models based on Scale, Shift and Rotation-Invariant Graphical Feature
【24h】

Retrieval Method for 3D Object Models based on Scale, Shift and Rotation-Invariant Graphical Feature

机译:基于比例,位移和旋转不变图形特征的3D对象模型检索方法

获取原文
获取原文并翻译 | 示例
           

摘要

We propose a new retrieval method for 3D object models. The method is based on our scale, shift and rotation-invariant graphical feature defined on every pair of polygons in each of the polygonal object. The graphical feature shows the angle between the polygon pair, the relative distance between the gravity points and the relative gross area of the polygon pairs with each angle and distance. Our graphical feature is also robust to the variations of preciseness of polygonal approximation as well as size of the polygonal patch. We have examined our feature in 3D object model retrieval from 2,500 objects. The precision ratio and recall ratios are 71.5% and 61.4% for the best twenty candidates, respectively.
机译:我们提出了一种新的3D对象模型检索方法。该方法基于我们在每个多边形对象中每对多边形上定义的缩放,平移和旋转不变的图形特征。图形功能显示了多边形对之间的角度,重力点之间的相对距离以及多边形对的相对总面积以及每个角度和距离。我们的图形功能对于多边形逼近的精度变化以及多边形补丁的大小也具有鲁棒性。我们已经检查了从2500个对象中检索3D对象模型的功能。最好的二十个候选者的准确率和召回率分别为71.5%和61.4%。

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