Motion Candidate Reordering for DMVR Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face challenges in improving coding efficiency, particularly in the reordering and management of motion candidate lists, which affects the overall compression and decoding processes.

Innovation Solution

The proposed methods involve reordering motion candidate lists through refined motion information, updating motion candidate lists by removing redundant candidates, and applying adaptive decoder side motion vector refinement to enhance coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If motion candidate lists are reordered using refined motion information, then coding efficiency is improved, but processing complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprocessing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by performing motion candidate list reordering and refinement operations during the encoding process before actual video coding. The motion candidate lists are reordered using refined motion information generated in advance, allowing the main coding process to benefit from optimized candidate arrangements without increasing real-time processing complexity during video decoding.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes parameters related to motion candidate list management by introducing reordering operations based on refined motion information. The motion candidates are rearranged according to their refined motion vectors, transforming the list from a conventional order to an optimized order that improves coding efficiency while maintaining manageable processing complexity through systematic parameter modification.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If motion candidate lists are updated by removing redundant candidates, then coding gain increases, but processing time increases

Engineering Contradiction:
Improvecoding gainVSAvoidprocessing time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent applies the extraction principle by identifying and removing redundant motion candidates from the candidate lists. Redundant candidates that do not contribute to improved coding performance are extracted and eliminated, resulting in a streamlined list of meaningful motion candidates that increases coding gain while the removal operation itself is performed efficiently during the encoding process.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the composition parameter of motion candidate lists by dynamically updating them to include only non-redundant candidates. This parameter modification optimizes the list content based on motion information analysis, removing unnecessary entries that would waste processing time during subsequent coding operations while maintaining the essential motion prediction capabilities.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If adaptive decoder side motion vector refinement is applied, then coding efficiency improves, but computational complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidcomputational complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies feedback by implementing adaptive decoder side motion vector refinement that uses decoded motion information to adjust and optimize motion vectors during the decoding process. The system feeds back refined motion vectors to improve prediction accuracy, enhancing coding efficiency while the adaptive nature allows the complexity to be managed through intelligent control of the refinement process.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20260019565A1Method, apparatus, and medium for video processing
Publication Date: 2026.01.15 BYTEDANCE INC
  • US20260019565A1 patent drawing
  • US20260019565A1 patent drawing
  • US20260019565A1 patent drawing

AI summary

Embodiments of the disclosure provide a solution for video processing. A method for video processing is proposed. The method includes: applying, during a conversion between a target block of a video and a bitstream of the video, a reordering process using refined motion information to the target block that is coded with an inter mode with a decoder side motion vector refinement (DMVR), to reorder at least one of a first motion candidate list or a second motion candidate list, wherein a first set of motion candidates in the first motion candidate list are refined in one direction and a second set of motion candidates in the second motion candidate list are refined in an-other direction; and performing the conversion based on the at least one of the reordered first motion candidate list or the reordered second motion candidate list.