• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Vision-based head pose estimation and interactivity analysis: Algorithms, systems and evaluation.
【24h】

Vision-based head pose estimation and interactivity analysis: Algorithms, systems and evaluation.

机译：基于视觉的头部姿势估计和交互性分析：算法，系统和评估。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Advances in visual computing have reached a point where systems can observe and interpret human activity. Learning to accomplish any human monitoring task needs a careful consideration of the level of observational detail that is required to understand the activity. In the general context of observing human interaction, much information can be obtained by looking at the head. The direction of a person's head gives a strong indication on what person or object they are focusing their attention. Since humans primarily attend to items of interest, estimating head pose can essentially be viewed as applying a saliency filter to determine the level of importance of each object in the field of view.;This dissertation presents new computational algorithms and systems for detecting and tracking people using head location, orientation, and movement as the salient cue. Head movement is used to monitor the interaction between people in 3D environment, and experimental evaluations are presented for both monocular and distributed camera setups in intelligent automobiles and meeting spaces.
机译：视觉计算的进步已经达到系统可以观察和解释人类活动的地步。学习完成任何人类监视任务都需要仔细考虑了解活动所需的观察细节水平。在观察人类交互的一般情况下，通过观察头部可以获得很多信息。一个人的头部方向可以清楚地表明他们正在关注哪个人或物体。由于人类主要关注感兴趣的物体，因此估计头的姿势本质上可以看作是应用显着性滤波器来确定视野中每个物体的重要性水平。本论文提出了用于检测和跟踪人的新的计算算法和系统。使用头部的位置，方向和移动作为显着提示。头部运动用于监视3D环境中人与人之间的互动，并针对智能汽车和会议空间中的单眼和分布式摄像头设置进行了实验评估。

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