发明公开
EP2492870A3 Method and system for generating a 3D representation of a dynamically changing 3D scene
失效 - 权利终止
转让

基本信息:
- 专利标题: Method and system for generating a 3D representation of a dynamically changing 3D scene
- 专利标题(中):方法和系统,用于产生动态变化的3D场景的3D表示
- 申请号:EP12002185.2 申请日:2007-05-24
- 公开(公告)号:EP2492870A3 公开(公告)日:2013-01-23
- 发明人: Würmlin Stadler, Stephan , Niederberger, Christoph
- 申请人: LiberoVision AG
- 申请人地址: Technoparkstrasse 1 8005 Zürich CH
- 专利权人: LiberoVision AG
- 当前专利权人: VIZRT AG
- 当前专利权人地址: VIZRT AG
- 代理机构: Frei Patent Attorneys
- 优先权: EP06405248 20060602
- 主分类号: G06T7/20
- IPC分类号: G06T7/20 ; G06T5/00
摘要:
A method for filling holes (802) that are left in a background of an image (201) after cutting out objects from the background comprises the steps of
• acquiring (102) an image (201) from a camera (702) observing a 3D scene (701);
• segmenting the image (201) by separating the objects from the background of the image (201);
• filling corresponding holes (802) left in the background by image inpainting, by mapping image data that corresponds to source patches (804a, 804b, 804c, 804d) comprising real image data to destination patches (803a, 803b, 803c, 803d) comprising unknown parts of the image, thereby filling the holes (802).
Therein, the mapping involves a transformation of the patches according to their spatial, that is, 3D relationship.
摘要(中):
• acquiring (102) an image (201) from a camera (702) observing a 3D scene (701);
• segmenting the image (201) by separating the objects from the background of the image (201);
• filling corresponding holes (802) left in the background by image inpainting, by mapping image data that corresponds to source patches (804a, 804b, 804c, 804d) comprising real image data to destination patches (803a, 803b, 803c, 803d) comprising unknown parts of the image, thereby filling the holes (802).
Therein, the mapping involves a transformation of the patches according to their spatial, that is, 3D relationship.
用于产生动态变化的三维场景的表示的方法,包括的€¢以下步骤从位于不同位置处的至少两个相机获取至少两个同步的视频流(120)和观察相同的3D场景(102); €¢确定性采矿摄像机参数,其包括取向和变焦设置,对于所述至少两个照相机(103); €¢跟踪物体的运动(310A,B,312A,B; 330A,B,331A,B,332A,B; 410A,B,411A,B; 430A,B,431A,B; 420A,B,421A, b)在所述至少两个视频流(104); €¢确定性采说的身份在所述至少两个视频流(105)对象; €¢确定性采矿通过合并从所述至少两个视频流(106)的信息中的对象的3D位置; worin跟踪的步骤(104)在所述至少两个视频流物体的运动使用来自对象的3D位置中的一个或多个较早的时刻在时间上导出的位置信息。 其结果,质量,速度和在视频流中的2D跟踪的鲁棒性得以改善。
公开/授权文献:
IPC结构图谱:
G | 物理 |
--G06 | 计算;推算;计数 |
----G06T | 一般的图像数据处理或产生 |
------G06T7/00 | 图像分析,例如从位像到非位像 |
--------G06T7/20 | .运动分析 |