×

Motion vector prediction for affine motion models in video coding

  • US 10,448,010 B2
  • Filed: 10/04/2017
  • Issued: 10/15/2019
  • Est. Priority Date: 10/05/2016
  • Status: Active Grant
First Claim
Patent Images

1. A method of decoding video data, the method comprising:

  • selecting a source affine block, the source affine block being an affine-coded block that spatially neighbors a current block;

    extrapolating motion vectors of a plurality of control points of the source affine block to determine motion vector predictors for a plurality of control points of the current block, wherein the control points of the source affine block include a first control point at a top-left of the source affine block and a second control point at a top-right of the source affine block, and the motion vectors of the first and second control points of the source affine block are used for affine prediction of the source affine block;

    inserting, into an affine motion vector predictor (MVP) set candidate list, an affine MVP set that includes the motion vector predictors for the control points of the current block;

    determining, based on an index signaled in a bitstream, a selected affine MVP set in the affine MVP set candidate list;

    obtaining, from the bitstream, motion vector differences (MVDs) that indicate differences between motion vectors of the control points of the current block and motion vector predictors in the selected affine MVP set;

    determining, based on the motion vector predictors included in the selected affine MVP set and the MVDs, motion vectors of the control points of the current block;

    generating, based on the motion vectors of the control points of the current block, a predictive block; and

    reconstructing the current block based on residual data and the predictive block.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×