Neural Reference Block Enhancement for Video Motion Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current interpolation filtering prediction methods in video coding exhibit low prediction efficiency, leading to poor video coding and decoding performance.

Innovation Solution

Perform quality enhancement on reference blocks using a neural network to obtain enhancement blocks, which are then used to determine prediction blocks, reducing decoding complexity and improving accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If current interpolation filtering prediction method is used, then video compression is achieved, but prediction efficiency is low resulting in poor coding and decoding performance

Engineering Contradiction:

Improveprediction accuracyVSAvoidcoding and decoding performance

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent introduces an enhancement block as an intermediary between the reference block and the prediction block. This enhancement block is generated by performing quality enhancement on the reference block, and then the prediction block is determined based on this enhanced block rather than directly from the original reference block. This intermediary structure improves prediction accuracy while maintaining coding efficiency

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent performs quality enhancement on the reference block in advance before generating the prediction block. This preliminary action of enhancing the reference block quality allows the subsequent prediction process to work with higher quality data, improving overall prediction efficiency without adding complexity to the core prediction mechanism

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260075257A1Decoding method, encoding method, training method, decoder, and encoder

Publication Date: 2026.03.12 GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD

AI summary

Embodiments of the present disclosure provide a decoding method, which includes: decoding a bitstream to determine a motion parameter of a current block; determining a reference block of the current block in a reference picture of the current block based on the motion parameter, performing quality enhancement on the reference block to obtain an enhancement block; and determining a prediction block of the current block based on the enhancement block.