【24h】

Robust and Active Human Face Tracking Vision Using Multiple Information

机译:使用多种信息的健壮主动的人脸跟踪视觉

获取原文
获取原文并翻译 | 示例

摘要

Up to now, a variety of methods for facial recognition has been proposed. Although most of them use still images taken under the restricted conditions, the more robust and more flexible recognition is required to apply for robot vision, man-machine interface, and security systems. In this paper we present an active tracking vision that can detect human faces robustly from sequential images under practical conditions. Our system efficiently recognizes a face based on the multiple information from simple agents that extract different image features. The agents are classified into two categories: real-time agents to respond quickly and semi real-time agents to perform rather complex processings, all of which interchange their respective results in order to achieve quick and reliable tracking using the CCD camera mounted on the manipulator. The experiments show that the proposed system could track some persons in the room and storage detected face images in the database file with the timestamps and position data automatically.
机译:迄今为止,已经提出了多种用于面部识别的方法。尽管它们中的大多数使用在受限条件下拍摄的静止图像,但是需要更强大,更灵活的识别才能应用于机器人视觉,人机界面和安全系统。在本文中,我们提出了一种主动跟踪视觉,可以在实际条件下从连续图像中可靠地检测人脸。我们的系统基于来自提取不同图像特征的简单代理的多种信息,有效地识别人脸。代理分为两类:快速响应的实时代理和执行相当复杂的处理的半实时代理,它们全部交换各自的结果,以便使用安装在操纵器上的CCD摄像机实现快速可靠的跟踪。 。实验表明,所提出的系统可以跟踪房间中的某些人,并将检测到的面部图像自动存储在带有时间戳和位置数据的数据库文件中。

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