首页> 外文期刊>電子情報通信学会技術研究報告. パターン認識·メディア理解. Pattern Recognition and Media Understanding >Improved method of image registration between frames and panoramic images using projective transform: for panorama-based annotation overlay
【24h】

Improved method of image registration between frames and panoramic images using projective transform: for panorama-based annotation overlay

机译:使用投影变换的帧和全景图像之间的图像配准的改进方法:用于基于全景的注释叠加

获取原文
获取原文并翻译 | 示例
           

摘要

In this paper, we propose an improved method of annotation overlay on live video which is one of the most promising applications in augmented reality and wearable computer. In our previous work, we proposed the panorama-based method of annotation overlay which aligns an input video frame and a panoramic image using affine transform. However, affine transform is generally not capable of image registration between a frame and a panorama. We improved the previous method so that it can estimate projective transform parameters without severely increasing computational cost. We confirmed that improved method could estimate image registration parameters under conditions that hindered the previous method. Its computational cost increased only by 10--20% and its real-time processing with software implementation was achieved.
机译:在本文中,我们提出了一种改进的实时视频注释叠加方法,这是增强现实和可穿戴计算机中最有前途的应用之一。在我们以前的工作中,我们提出了基于全景图的注释叠加方法,该方法使用仿射变换将输入视频帧和全景图像对齐。但是,仿射变换通常无法在帧和全景之间进行图像配准。我们改进了先前的方法,以便它可以估计投影变换参数而不会严重增加计算成本。我们确认,改进的方法可以在阻碍前一种方法的条件下估计图像配准参数。它的计算成本仅增加了10--20%,并且实现了带有软件实现的实时处理。

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