首页> 外文会议>Biomedical Engineering >MARKER SYSTEM FOR VISUAL IDENTIFICATION OF HUMAN KINEMATICS
【24h】

MARKER SYSTEM FOR VISUAL IDENTIFICATION OF HUMAN KINEMATICS

机译:视觉识别人类运动学的标记系统

获取原文

摘要

Until today different commercial systems in visual identification of human kinematics were developed, like VICON's Workstation, Ariel system or others. Principal problem relies in complicated structure and variable pose of human body resulting in an often self-occlusion of particular segments. Visual identification using markers placed on crucial positions of the body (like joints) for the specific motion observed is matter of many actual investigations. Tracking of markers is still a problem. Principal reasons are the occlusion of markers and noise introduced capturing video sequences of motion. In this paper we describe a low-cost system that reduces self-occlusion and other errors in visual identification of human kinematics using markers. Also there is a brief description of the software "Marker Tracking" that was realized as the central part of this system. Testing showed that this software is reliable and accurate, depending only on the quality of the equipment used for video registration. Principal scope of this paper is to describe this system from the users point of view and to give introduction to the methods used in the software.
机译:直到今天,人们仍在开发各种用于视觉识别人体运动学的商业系统,例如VICON的工作站,Ariel系统或其他系统。主要问题在于人体复杂的结构和可变的姿势,导致特定部分经常自我闭塞。对于观察到的特定运动,使用放置在人体关键部位(如关节)的标记进行视觉识别是许多实际研究的重点。跟踪标记仍然是一个问题。主要原因是标记的遮挡和引入的噪声捕获了运动的视频序列。在本文中,我们描述了一种低成本的系统,该系统可减少使用标记物进行的人体运动学视觉识别中的自我遮挡和其他错误。此外,还对软件“标记跟踪”进行了简要说明,该软件已实现为该系统的核心部分。测试表明,该软件可靠且准确,仅取决于用于视频注册的设备的质量。本文的主要范围是从用户角度描述该系统,并介绍该软件中使用的方法。

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