Video error concealment method
Disclosed herein is a system and method for reconstruction video information lost as a result of transmission errors. The system and method have four aspects, including: (1) changing the bit and/or packet rate; (2) inserting redundant information into the video bitstream; (3) providing automatic refresh of certain regions of the video on a periodic basis; and (4) interleaving coded macroblocks into diversity groups for transmission to spatially spread the effect of lost packets. The image reconstruction may then take advantage of these three aspects to provide an enhanced result in the presence of transmission losses.
1. A method of digitally encoding a video bitstream to facilitate the concealment of errors caused by loss of some portion of the video bitstream, the method comprising:
inserting redundant information in the video bitstream on an ongoing basis, wherein the redundant information comprises a map specifying whether each block was coded using intra or inter coding and redundant motion vectors for each inter coded block; and
overlapping the redundant information over a plurality of packets such that at least one packet contains redundant information corresponding to two or more previous packets,
wherein the redundant information is inserted into the bitstream at the request of a receiver.
2. The method of claim 1 wherein the redundant information is inserted into the bitstream by an MCU.
3. The method of claim 1 further comprising periodically coding each block of the video bitstream as an intra coded block using a deterministic pattern.
4. The method of claim 3 wherein the deterministic pattern is a walk-around refresh.
5. The method of claim 1 further comprising coding each block of the video bitstream determined to be in a motion wake as an intra coded block.
6. The method of claim 1 further comprising interleaving coded macroblocks into a plurality of diversity groups for transmission, wherein spatially adjacent macroblocks are assigned to different diversity groups.
7. The method of claim 1 further comprising assigning each of a plurality of macroblocks to one of a plurality of diversity groups such that each diversity group contains non-spatially adjacent macroblocks, wherein the diversity groups are transmitted in separate packets.
8. A method of decoding a video bitstream, the method comprising:
receiving a video bitstream wherein at least some portions of the video bitstream are lost;
identifying at least one lost macroblock; and
referring to a redundant inter macroblock map embedded in the video stream by an MCU at the request of a receiver, in order to determine how to reconstruct the lost macroblock,
wherein if the inter macroblock map indicates that the lost macroblock was inter coded, the lost macroblock is reconstructed using redundant motion vectors embedded in the video bitstream by an MCU at the request of a receiver.
9. The method of claim 8 wherein if the inter macroblock map indicates that the lost macroblock was intra coded, the lost macroblock is reconstructed as an uncoded macroblock.
10. The method of claim 8 wherein if the inter macroblock map indicates that the lost macroblock was intra coded, the lost macroblock is reconstructed by spatial interpolation of adjacent macroblocks.
11. A method of decoding a video bitstream, the method comprising:
receiving a video bitstream wherein at least some portions of the video bitstream are lost;
identifying at least one lost macroblock; and
referring to a redundant inter macroblock map embedded in the video stream by an MCU at the request of a receiver, to determine how to reconstruct the lost macroblock,
wherein if the inter macroblock map indicates that the lost macroblock was intra coded and the macroblock is at an edge of a picture, the macroblock is not reconstructed.
12. The method of claim 11 wherein if the inter macroblock map indicates that the lost macroblock was inter coded, the lost macroblock is reconstructed by estimating motion vectors from one or more surrounding macroblocks.
13. The method of claim 12 wherein estimating motion vectors from one or more surrounding macroblocks comprises computing the median of motion vectors of a plurality of neighboring macroblocks.
14. A method of conducting a multipoint video conference wherein a multipoint control unit facilitates the concealment of errors on at least one link caused by loss of some portion of a video bitstream along that link, the method comprising:
inserting redundant information into the video bitstream on an ongoing basis, wherein the redundant information comprises a map specifying whether each block was coded using intra or inter coding and redundant motion vectors for each inter coded block; and
overlapping the redundant information over a plurality of packets such that at least one packet contains redundant information corresponding to two or more previous packets.