Scalable video encoding in a multi-view camera system

摘要

The present invention employs scalable video coding (SVC) in a multi-view camera system, which is particularly suited for video conferencing. Multiple cameras are oriented to capture video content of different image areas and generate corresponding original video streams that provide video content of the image areas. An active one of the image areas may be identified at any time by analyzing the audio content originating from the different image areas and selecting the image area that is associated with the most dominant speech activity.

主权项

1. A method comprising:
receiving a video steam from each of a plurality of cameras at a conference location; identifying at least one of the video streams as active for the conference location; encoding the at least one of the video streams at a first quality level using scalable video coding (SVC) based on said identifying, wherein said encoding the at least one of the video streams produces a first one or more encoded video streams; encoding others of the video streams at a second quality level using SVC based on said identifying, wherein the second quality level is lower than the first quality level, and wherein said encoding the others of the video streams produces a second one or more encoded video streams; and providing the first one or more encoded video streams and the second one or more encoded video streams to at least one device outside of the conference location.