Video Coding Reference Picture Selection Across Random Access Segments

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional video coding technologies limit video coding/decoding efficiency by restricting inter prediction across random access segments, which hampers the use of reference pictures from outside the segment for encoding and decoding.

Innovation Solution

A method and apparatus that allow selecting reference pictures from a knowledge base for encoding and decoding, where at least one reference picture does not belong to the same random access segment as the current picture, expanding the candidate reference picture range by including pictures from adjacent segments based on similarity and decoding order.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If inter prediction across random access segments is restricted, then random access function is maintained, but video coding efficiency deteriorates

Engineering Contradiction:
Improverandom access functionVSAvoidvideo coding efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The video stream is divided into multiple random access segments with clear boundaries, allowing independent decoding within each segment while enabling controlled reference across segments through the segment boundary mechanism

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces segment boundary pictures as intermediary elements that mediate between segments, allowing reference picture lists to include pictures from other segments only when appropriate boundary conditions are met, thus enabling cross-segment prediction while maintaining random access capability

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If reference picture range is limited to same segment, then decoding complexity is reduced, but coding efficiency deteriorates

Engineering Contradiction:
Improvedecoding complexityVSAvoidcoding efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The reference picture selection mechanism is made dynamic: the decoder adaptively determines whether cross-segment reference is permitted based on picture type, segment boundary position, and reference picture list configuration, optimizing between complexity and efficiency for each coding scenario

Inventive Principle:
Principle #15Dynamics

3Adaptability or versatility

If cross-segment reference is allowed, then coding flexibility is improved, but random access capability deteriorates

Engineering Contradiction:
Improvecoding flexibilityVSAvoidrandom access capability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

Different reference picture selection rules are applied locally within different segment contexts: intra-segment reference is always permitted, while inter-segment reference is conditionally permitted based on local picture characteristics and segment boundary positions, maintaining random access at segment boundaries

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20240129455A1Picture encoding/decoding method and related apparatus
Publication Date: 2024.04.18 HUAWEI TECH CO LTD
  • US20240129455A1 patent drawing
  • US20240129455A1 patent drawing
  • US20240129455A1 patent drawing

AI summary

A picture encoding/decoding method and a related apparatus are provided. The picture decoding method includes obtaining a current picture from a video, the current picture being segmented using random access points into multiple video segments supporting a random access function; selecting, from a knowledge base, K reference pictures of the current picture, wherein at least one reference picture in the knowledge base does not belong to a random access segment in which the current picture is located, the random access segment comprising a picture sequence arranged in a decoding order from a closest random access point before the current picture to a closest random access point after the current picture; and decoding the current picture according to the K reference pictures.