Method and device for pruning a video sequence

A non-transitory computer storage medium having stored thereon instructions for implementing a method, when executed on a device having processing capabilities, of pruning a prediction-coded video sequence, the method comprising:

• a) obtaining one or more global motion values for the video sequence, wherein each global motion value represents an amount of motion in a scene depicted by a set of successive image frames in the video sequence;
• b) for a set of successive image frames having a global motion value being below a global motion threshold, identifying a group of pictures (GOP) that immediately follows the set of successive image frames in the video sequence, wherein the identified GOP comprises a sub-set of image frames of the video sequence;
• c) determining a frame configuration of the image frames of the identified GOP, wherein the determined frame configuration defines settings of frame parameters that need to be the same for a replacement frame in the identified GOP as for the rest of the frames in the identified GOP;
• d) obtaining a skip frame having the determined frame configuration, wherein the skip frame is an inter frame that directly references existing image data of at least one previously coded frame of the video sequence and that is coded without interpolation of residual information;
• e) replacing the initial intra frame of the identified GOP with the obtained skip frame;
• and
• f) configuring the skip frame and the remaining image frames of the identified GOP with a frame numbering that continues a frame numbering of the GOP immediately preceding the identified GOP, whereby a new GOP is formed of frames of the immediately preceding GOP and the identified GOP,
• wherein replacing the initial intra frame of the identified GOP with the obtained skip frame and configuring the skip frame and the remaining image frames of the identified GOP with a frame numbering that continues a frame numbering of the GOP immediately preceding the identified GOP are performed without transcoding the remaining image frames of the identified GOP, including without decoding and re-encoding the remaining image frames of the identified GOP.

A device for pruning a prediction-coded video sequence, the device comprising a processing device being adapted to:

• obtain one or more global motion values for the video sequence, wherein each global motion value represents an amount of motion in a scene depicted by a set of successive image frames in the video sequence,
• for a set of successive image frames having a global motion value being below a global motion threshold, identify a group of pictures (GOP) that immediately follows the set of successive image frames in the video sequence, wherein the identified GOP comprises a sub-set of image frames of the video sequence,
• determine a frame configuration of the image frames of the identified GOP, wherein the determined frame configuration defines settings of frame parameters that need to be the same for a replacement frame in the identified GOP as for the rest of the frames in the identified GOP,
• obtain a skip frame having the determined frame configuration, wherein the skip frame is an inter frame that directly references existing image data of at least one previously coded frame of the video sequence and that is coded without interpolation of residual information,
• replace the initial intra frame of the identified GOP with the obtained skip frame, and
• configure the skip frame and the remaining image frames of the identified GOP with a frame numbering that continues a frame numbering of the GOP immediately preceding the identified GOP, whereby a new GOP is formed of frames of the immediately preceding GOP and the identified GOP,
• wherein replacing the initial intra frame of the identified GOP with the obtained skip frame and configuring the skip frame and the remaining image frames of the identified GOP with a frame numbering that continues a frame numbering of the GOP immediately preceding the identified GOP are performed without transcoding the remaining image frames of the identified GOP, including without decoding and re-encoding the remaining image frames of the identified GOP.