Reference Picture List Signaling for Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding standards face challenges in efficiently signaling and managing reference pictures for temporal prediction, particularly in HEVC, where the complexity of reference picture lists can lead to increased bitrates and computational overhead.

Innovation Solution

The proposed solution involves improving the signaling schemes and construction processes for reference picture lists in HEVC, specifically by generating ordered lists L0 and L1, and using a unified signaling syntax to modify these lists efficiently, thereby reducing signaling overhead and improving prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple reference pictures are used for temporal prediction, then prediction accuracy is improved, but signaling overhead and computational complexity increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidreference picture list complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The reference picture list is segmented into two separate lists (L0 and L1) instead of using a single unified list. This segmentation allows the encoder to selectively manage and signal references from each list independently, reducing the overall signaling overhead while maintaining access to multiple reference pictures for improved prediction accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic modification mechanisms for reference picture lists, allowing the number and composition of references in L0 and L1 to be adaptively adjusted based on picture type (I, P, B pictures) and coding conditions. This dynamic approach enables the system to optimize between using more references for accuracy versus using fewer references to reduce complexity, depending on the specific coding scenario.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If multiple reference pictures are used for temporal prediction, then prediction accuracy is improved, but bitrate increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidsignaling overhead
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

By dividing the reference picture list into L0 and L1, the patent enables more efficient signaling where the decoder only needs to receive indicators for the specific list being used (L0 for P pictures, L1 for B pictures), rather than signaling all possible references. This segmentation reduces the amount of information that needs to be transmitted while still providing access to multiple reference pictures for accurate prediction.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameter structure by introducing separate list identifiers (L0/L1) and using picture type (I/P/B) as a controlling parameter to determine which list to use. This parameter-based approach allows the system to switch between different reference configurations without transmitting detailed information about each reference picture, thereby reducing signaling overhead while maintaining prediction accuracy.

Inventive Principle:
Principle #35Parameter changes

3Adaptability or versatility

If reference picture lists are modified frequently, then adaptation to different coding conditions is improved, but computational complexity increases

Engineering Contradiction:
Improveadaptation to coding conditionsVSAvoidcomputational complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent uses picture type (I, P, B) as a simple controlling parameter to determine the reference picture list configuration. This parameter-based approach allows frequent adaptation to different coding conditions without requiring complex real-time analysis, as the picture type itself directly indicates which reference list (L0 or L1) should be used and how many references to employ.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The reference picture lists L0 and L1 are pre-configured and maintained in the decoded picture buffer before actual prediction occurs. This preliminary organization of references allows the encoder and decoder to quickly adapt to different coding conditions by simply selecting from the pre-prepared lists based on picture type, avoiding the need for complex real-time list construction and reducing computational complexity during the actual coding process.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20250150577A1Method and apparatus for signaling and construction of video coding reference picture lists
Publication Date: 2025.05.08 INTERDIGITAL VC HOLDINGS INC
  • US20250150577A1 patent drawing
  • US20250150577A1 patent drawing
  • US20250150577A1 patent drawing

AI summary

Improved method and apparatus for signaling of reference pictures used for temporal prediction. The signaling schemes and construction process for different reference picture lists in HEVC Working Draft 5 (WD5) are improved.