Selection of pictures for disparity vector derivation

摘要

When coding multiview video data, a video encoder and video decoder may select a candidate picture from one of one or more random access point view component (RAPVC) pictures and one or more pictures having a lowest temporal identification value. The video encoder and video decoder may determine whether a block in the selected candidate picture is inter-predicted with a disparity motion vector and determine a disparity vector for a current block of a current picture based on the disparity motion vector. The video encoder and video decoder may inter-prediction encode or decode, respectively, the current block based on the determined disparity vector.

主权项

1. A method for decoding a bitstream that includes multiview video data that includes multiple views, the method comprising:
applying a neighboring block based disparity vector (NBDV) generation process for purposes of deriving a disparity vector for a current block of a current picture of the video data, wherein the NBDV generation process relies on motion vectors of neighboring blocks for deriving the disparity vector for the current block of the current picture relative to another view of the multiview video data, wherein applying the NBDV generation process comprises:
determining whether one or more random access point view component (RAPVC) pictures are available;selecting a candidate picture from the one or more RAPVC pictures based on the determination that at least one of the RAPVC pictures is available, orselecting the candidate picture from one or more pictures having a lowest temporal identification value based on the determination that no RAPVC picture is available;determining whether a video block in the selected candidate picture is inter-predicted with a disparity motion vector, wherein the disparity motion vector refers to a block in a view other than a view that includes the selected candidate picture of the multiview video data; anddetermining the disparity vector for the current block of the current picture, as part of the NBDV generation, based on the disparity motion vector based on the video block in the selected candidate picture of the multiview video data being inter-predicted with the disparity motion vector; and inter-prediction decoding the current block based on the disparity vector.