Compressed video signal including information for independently coded regions

Abstract

In a MPEG or other video system, a 'region map' is preferably defined in header information for each frame which indicates for each independently coded region in the frame which 'image slices' contain data for that region. Using a system of globally and locally defined regions (region groups), the map permits automatic identification of an object through multiple frames (notwithstanding movement) and for select decoding and extraction of that object only, without necessarily decoding all irrelevant image data. An edited region can be re-inserted into the original encoded bit stream in place of the extracted data using minimal computational resources. Using this signal format, objects or locations within a video sequence can be organized into region groups such that they can be tracked from frame to frame and associated with each other; each region group is independently coded, such that motion vectors and residuals of a dependent frame (such as a MPEG “P” or “B” frame) point only to a corresponding region in an anchor frame.