• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Twentieth International Joint Conference on Artificial Intelligence(IJCAI-07) >MESH-based Active Monte Carlo Recognition (MESH-AMCR)
【24h】

MESH-based Active Monte Carlo Recognition (MESH-AMCR)

机译：基于MESH的主动蒙特卡洛识别（MESH-AMCR）

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 we extend Active Monte Carlo Recognition (AMCR), a recently proposed framework for object recognition. The approach is based on the analogy between mobile robot localization and object recognition. Up to now AMCR was only shown to work for shape recognition on binary images. In this paper, we significantly extend the approach to work on realistic images of real world objects. We accomplish recognition under similarity transforms and' even severe non-rigid and non-affine deformations. We show that our approach works on databases with thousands of objects, that it can better discriminate between objects than state-of-the art approaches and that it has significant conceptual advantages over existing approaches: It allows iterative recognition with simultaneous tracking, iteratively guiding attention to discriminative parts, the inclusion of feedback loops, the simultaneous propagation of multiple hypotheses, multiple object recognition and simultaneous segmentation and recognition. While recognition takes place triangular meshes are constructed that precisely define the correspondence between input and prototype object, even in the case of strong non-rigid deformations.
机译：在本文中，我们扩展了主动蒙特卡洛识别（AMCR），这是最近提出的对象识别框架。该方法基于移动机器人定位和对象识别之间的类比。到目前为止，AMCR仅显示可用于二进制图像的形状识别。在本文中，我们极大地扩展了在现实世界对象的真实图像上工作的方法。我们在相似变换甚至严重的非刚性和非仿射变形下完成识别。我们证明了我们的方法适用于具有数千个对象的数据库，与现有方法相比，它可以更好地区分对象，并且与现有方法相比，它具有显着的概念优势：它允许迭代识别，同时跟踪，迭代地引导注意力到区分部分，包括反馈环，多个假设的同时传播，多个对象识别以及同时的分段和识别。进行识别时，即使在强烈的非刚性变形的情况下，也会构造三角形网格，以精确定义输入对象与原型对象之间的对应关系。

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