Method and apparatus for video coding using motion vector with component-wise adaptive spatial resolution
AssigneesHyundai Motor Company, Kia Corporation, Research & Business Foundation Sungkyunkwan University
InventorsByeung Woo Jeon, Bum Yoon Kim, Yu Jin Lee, Seung Wook Park
Life of this patent
Estimating the term…
Abstract
A video coding method and apparatus adaptively determine spatial resolution for a horizontal component and a vertical component of a motion vector or block vector indicating a position of a reference block when performing inter prediction or intra block copy (IBC) of a current block. The video coding method and apparatus separately determine each spatial resolution for a plurality of motion vectors.
Claims13
A method of decoding a current block coded in an affine prediction mode using a plurality of control points, performed by a video decoding apparatus, the method comprising:
- decoding, from a bitstream, an AMVR (Adaptive Motion Vector Resolution) flag, wherein the AMVR flag indicates whether or not adaptive spatial resolution is applied to motion vector differences of the plurality of control points;
- and
- in response to the AMVR flag indicating that the adaptive spatial resolution is applied,
- decoding, from the bitstream, a plurality of precision indices corresponding to the plurality of control points, respectively, and
- determining spatial resolution values of the motion vector differences of the plurality of control points based on the plurality of precision indices.
A method of encoding a current block coded in an affine prediction mode using a plurality of control points, performed by a video encoding apparatus, the method comprising:
- determining spatial resolution values of motion vector differences of the plurality of control points;
- determining an AMVR (Adaptive Motion Vector Resolution) flag based on the spatial resolution values, wherein the AMVR flag indicates whether or not adaptive spatial resolution is applied to the motion vector differences of the plurality of control points;
- encoding the AMVR flag into a bitstream;
- and
- in response to the AMVR flag indicating that the adaptive spatial resolution is applied, encoding, into the bitstream, a plurality of precision indices indicating the spatial resolution values of the motion vector differences of the plurality of control points, respectively.
A method for transmitting a bitstream containing encoded video data, the method comprising:
- generating the bitstream for a current block in an image coded in an affine prediction mode using a plurality of control points;
- and
- transmitting the bitstream,
- wherein generating the bitstream comprises:
- determining spatial resolution values of motion vector differences of the plurality of control points;
- determining an AMVR (Adaptive Motion Vector Resolution) flag based on the spatial resolution values, wherein the AMVR flag indicates whether or not adaptive spatial resolution is applied to the motion vector differences of the plurality of control points;
- encoding the AMVR flag into a bitstream;
- and
- in response to the AMVR flag indicating that the adaptive spatial resolution is applied, encoding, into the bitstream, a plurality of precision indices indicating the spatial resolution values of the motion vector differences of the plurality of control points, respectively.
Dependencies are inferred from the claim text.