Video coding methods and apparatuses performed by a video encoder and a decoder

By constructing a candidate list for affine motion vector prediction and utilizing the combination of spatial and temporal MV, the problem of low efficiency of affine motion models in video encoding and decoding in existing technologies is solved, achieving more efficient encoding and decoding performance.

CN116708814BActive Publication Date: 2026-06-05HFI INNOVATION INC
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HFI INNOVATION INC
Filing Date
2019-05-24
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Existing video encoding and decoding technologies are inefficient when handling complex motion and struggle to effectively utilize affine motion models for encoding and decoding, resulting in insufficient encoding and decoding performance.

Method used

A candidate list for affine motion vector prediction is constructed by using multiple spatially adjacent blocks and neighboring blocks of co-located blocks. A affine MVP candidate is derived through a four-parameter or six-parameter affine model and encoded and decoded in a merging mode. Prediction is then performed by combining spatial and temporal MV.

Benefits of technology

It improves the efficiency and performance of video encoding and decoding, especially when dealing with complex motion scenes, reducing redundant information and improving encoding and decoding quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116708814B_ABST
    Figure CN116708814B_ABST
Patent Text Reader

Abstract

Methods and apparatuses for coding a system using an affine motion model are disclosed. According to one method, a neighboring block set of a current block including a plurality of spatial neighboring blocks and one or more collocated blocks is determined for the current block. One or more constructed affine MVP (motion vector prediction) candidates are derived for an affine MVP candidate list based on control point (CP) motion vectors (MVs) at the plurality of spatial neighboring blocks and the one or more collocated blocks. The constructed affine MVP candidate without a temporal MV is checked and inserted into the affine MVP candidate list before any constructed affine MVP with a temporal MV. Then, the motion information of the current block or the current block is encoded or decoded based on the affine MVP candidate list.
Need to check novelty before this filing date? Find Prior Art