| 摘要 |
Method and apparatus for capturing stereoscopic panoramic image of virtual environment comprising: pausing state of virtual environment s710; for each of a plurality (may be four, five, six or eight) of positions (350A-H, fig. 6) in virtual environment that form vertices of bounding space (360, fig. 6) within which left and right viewpoints of stereoscopic image are formed: rendering s720 and capturing s730 plural images encompassing panoramic view; capturing data for distance between position and surfaces or objects visible from that position s740. Image can be viewed by: accessing picture and range data (s810, fig. 8); defining locations and orientation (s820, fig. 8), and obtaining images (s830, fig. 8), for stereoscopic eye viewpoints within bounding space; generating spherical projection of images (s840, fig. 8). Pixels may be shifted based on distance and resulting gaps may be filled from other images. Image can be served to client device by: including data indicating at least two positions that do not use depth data; sending only images for those positions if device signals it will not use distance information (s830, fig. 9); sending all images for all positions if device will use distance (s840, fig. 9). Captured images may form a cube map or a dual paraboloid. Images may be distorted to correct for spherical projection. Imgaes may include more detail than real-time renders. Graphical elements such as game players hands or game menus may be omitted from renders. User interface control may activate image capture. |
