Automatic image alignment in video conferencing

摘要

In one implementation, at least two images are captured from a respective at least two cameras of a telepresence system. The at least two cameras have horizontally overlapping fields of view such that the at least two images horizontally overlap. A processor identifies, by image processing of the overlap of the at least two images, portions of each of the at least two images. The portions spatially correspond to adjacent displays and do not include the overlap. Subsequent images captured by the at least two cameras are displayed on the adjacent displays in a video conference. The displayed images are for the portions of the field of view corresponding to the adjacent displays.