Scaled Motion Vector Prediction for Reference List Mismatch

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding systems face challenges in efficiently predicting motion vectors, particularly when reference pictures are not included in the reference picture list, leading to suboptimal decoding and encoding processes.

Innovation Solution

The proposed solution involves obtaining a guiding motion vector associated with a reference picture not in the reference picture list, scaling it temporally based on calculated temporal distances, and incorporating it into the reference picture list for improved motion vector prediction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a guiding motion vector associated with a reference picture not in the reference picture list is used directly, then motion vector prediction can be performed, but the prediction accuracy is suboptimal because the reference picture is not available in the current reference picture list

Engineering Contradiction:
Improvemotion vector prediction accuracyVSAvoidreference picture list compatibility
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces a temporal scaling operation as an intermediary process that adapts the guiding motion vector from a reference picture outside the current reference picture list. The scaling factor, calculated based on temporal distances between pictures, acts as a mediator to transform the incompatible motion vector into a form that can be effectively used with the current reference picture list, thereby resolving the compatibility issue while preserving prediction accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the parameter of the guiding motion vector by applying temporal scaling. The motion vector is transformed using a scaling factor derived from temporal distance ratios, which adjusts the vector's magnitude to account for different time intervals between reference pictures. This parameter transformation enables the motion vector to remain accurate even when the reference picture is not in the current reference picture list

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If temporal scaling is applied to the guiding motion vector, then the scaled guiding motion vector becomes compatible with the reference picture list, but additional calculation steps are required

Engineering Contradiction:
Improvereference picture list compatibilityVSAvoidcalculation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary calculations of temporal distances and scaling factors before applying them to the guiding motion vector. By pre-computing the temporal distance between the current picture and reference pictures, and determining the scaling factor in advance, the actual motion vector scaling operation becomes a simple multiplication, reducing the complexity of the main decoding/encoding path while maintaining compatibility

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If the guiding motion vector is temporally scaled using temporal distances, then the scaled guiding motion vector achieves better prediction accuracy, but the encoding/decoding process requires additional processing steps

Engineering Contradiction:
Improvemotion vector prediction accuracyVSAvoidencoding/decoding efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent implements self-service by having both the encoder and decoder independently perform the same temporal scaling calculations using identical algorithms and reference picture information. Since both sides compute the scaling factor and apply it to the guiding motion vector in the same manner, no additional signaling is required in the bitstream, and the process becomes self-contained, improving efficiency by avoiding redundant transmissions

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4679821A1Scaled chained motion vector prediction candidate
Publication Date: 2026.01.14 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP4679821A1 patent drawingFigure 1A
  • EP4679821A1 patent drawingFigure 1B
  • EP4679821A1 patent drawingFigure 1C

AI summary

Systems, methods, and instrumentalities are disclosed for performing video coding using a scaled chain Motion Vector Predictor (MVP) candidate. A reference picture list may be obtained for a current block. A guiding motion vector may be obtained that is associated with the current block. The guiding motion vector may be associated with a reference picture. The reference picture associated with the guiding motion vector may be determined to not be in the reference picture list of the current block. Based on this determination, the guiding motion vector may be temporally scaled such that the scaled guiding motion vector is associated with a reference picture in the reference picture list of the current block. The current block may be decoded based on the scaled guiding motion vector.