• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>The international journal of virtual reality >Towards Achieving Robust Video Self-avatar under Flexible Environment Conditions
【24h】

Towards Achieving Robust Video Self-avatar under Flexible Environment Conditions

机译：努力在灵活的环境条件下实现强大的视频自头像

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

The user's sense of presence within a virtual environment is very important as it affects their propensity to experience the virtual world as if it were real. A common method of immersion is to use a head-mounted display (HMD) which gives the user a stereoscopic view of the virtual world encompassing their entire field of vision. However, the disadvantage to using an HMD is that the user's view of the real world is completely blocked including the view of his or her own body, thereby removing any sense of embodiment in the virtual world. Without a body, the user is left feeling that they are merely observing a virtual world, rather than experiencing it. We propose using a video-based see-thru HMD (VSTHMD) to capture video of the view of the real-world and then segment the user's body from that video and composite it into the virtual environment. We have developed a VSTHMD using commercial-off-the-shelf components, and have developed a preliminary algorithm to segment and composite the user's arms and hands. This algorithm works by building probabilistic models of the appearance of the room within which the VSTHMD is used, and the user's body. These are then used to classify input video pixels in real-time into foreground and background layers. The approach has promise, but additional measures need to be taken to more robustly handle situations in which the background contains skin-colored objects such as wooden doors. We propose several methods to eliminate such false positives, and discuss the initial results of using the 3D data from a Kinect to identify false positives.
机译：用户在虚拟环境中的存在感非常重要，因为它会影响他们体验虚拟世界的意愿，就像现实世界一样。一种常见的沉浸式方法是使用头戴式显示器（HMD），该显示器可为用户提供虚拟世界的立体视图，其中涵盖了他们的整个视野​​。但是，使用HMD的缺点是，用户对现实世界的视图（包括他或她自己的身体的视图）被完全遮挡，从而消除了虚拟世界中的任何体现感。没有身体，用户会感到自己只是在观察虚拟世界，而不是在体验虚拟世界。我们建议使用基于视频的透视HMD（VSTHMD）捕获真实世界视图的视频，然后从该视频中分割用户的身体并将其组合到虚拟环境中。我们使用现成的商用组件开发了VSTHMD，并开发了用于分割和合成用户手臂和手部的初步算法。该算法通过建立使用VSTHMD的房间外观以及用户身体的概率模型来工作。然后使用这些将输入视频像素实时分类为前景和背景层。该方法有望实现，但是需要采取其他措施来更稳健地处理背景中包含皮肤色对象（例如木门）的情况。我们提出了几种消除此类误报的方法，并讨论了使用Kinect的3D数据识别误报的初步结果。

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