Error concealment method in a motion video decompression system
AssigneeLG Electronics Inc.
InventorsEung-Tae Kim, Hee-Bok Park, Jae-Won Suh, Yo-Sung Ho
Life of this patent
Estimating the term…
Abstract
An error concealment method of a motion video decompression system is disclosed. The present invention uses motion vector estimation to recover video signals, with errors generated by a transmission state, nearly to the original video. Particularly, the present method includes estimating motion information of a lost video, and restoring the lost video from a video adjacent to the lost video by using the estimated motion information.
Claims20
An error concealment method of a motion video decompression system, comprising:
- estimating motion information of lost video data within a current video frame;
- and
- restoring the lost video data from a video frame adjacent to a current video frame utilizing the estimated motion information.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating a motion vector of the lost macroblock by the motion information of the top or bottom macroblock, if motion information exists in only one of either the top or bottom macroblock;
- and
- searching a macroblock of a previous frame adjacent to the video frame using the estimated motion vector, and recovering the lost macroblock from the searched macroblock of the previous frame.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating a motion vector of the lost macroblock by an average value between the motion information of the top and bottom macroblocks, if motion information exists in both the top and bottom macroblock;
- and
- searching a macroblock of a previous frame adjacent to the video frame using the estimated motion vector, and recovering the lost macroblock from the searched macroblock of the previous frame.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a bidirectional predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating a motion vector of the lost macroblock by an average value between motion information of the top and bottom macroblocks, if the motion information exists in both the top and bottom macroblock;
- and
- searching a macroblock of a previous frame adjacent to the video frame using the estimated motion vector, and recovering the lost macroblock from the searched macroblock of the previous frame.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a bidirectional predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating pixels of the lost macroblock by an average pixel value between pixels of macroblocks corresponding to the motion information of the top or bottom macroblocks, if motion information exists in both the top and bottom macroblock;
- and
- replacing the pixels of the lost macroblock with the average pixel value to recover the lost macroblock.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a bidirectional predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating a motion vector of the lost macroblock by the motion information of the top or bottom macroblock, if motion information exists in only one of either the top or bottom macroblock;
- and
- searching a macroblock of a previous frame adjacent to the video frame using the estimated motion vector, and recovering the lost macroblock from the searched macroblock of the previous frame.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a bidirectional predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating pixels of the lost macroblock by pixels of macroblocks corresponding to the motion information of the top or bottom macroblock, if motion information exists in only one of either the top or bottom macroblock;
- and
- replacing the pixels of the lost macroblock with the average pixel value to recover the lost macroblock.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock, wherein the determined type of video frame is a bidirectional predictive frame;
- detecting whether motion information exists in a top macroblock and a bottom macroblock of the lost macroblock;
- estimating motion information of lost video data within a current video frame;
- and
- recovering the information of the lost macroblock from the information of the macroblock positioned at the same location within the previous frame, if motion information does not exists in either the top or bottom macroblock.
An error concealment method of a motion video decompression system, comprising:
- determining a type of video frame containing a lost macroblock;
- estimating motion information of lost video data within a current video frame, wherein if one or more bidirectional predictive frame are within a Group of Picture, the motion information is estimated by assigning weights to forward motion information of a previous frame and to backward motion information of a subsequent frame, based upon the distance of the video frame from the previous frame and the subsequent frame;
- and
- recovering the information of the lost macroblock from the information of the macroblock positioned at the same location within the previous frame, if motion information does not exists in either the top or bottom macroblock.
Dependencies are inferred from the claim text.