Adaptive Video Resampling via Picture Order Count Constraints

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In video coding and decoding, especially for applications like 360 coding and surveillance, there is a need for adaptive resolution settings to manage different scene activities, as multiple semantically independent picture parts require separate resolution adjustments to optimize bandwidth and storage efficiency.

Innovation Solution

The method involves signaling adaptive picture size in a video bitstream by determining whether references are associated with access units based on picture order count (POC) signal values, allowing for different resampling factors for semantically independent pictures, and including POC signal values in the video parameter set to manage reference picture resampling and adaptive resolution coding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a fixed picture size is used for the entire coded video sequence, then device complexity is reduced and ease of operation is improved, but adaptability to different scene activities deteriorates

Engineering Contradiction:

Improveadaptability to scene activitiesVSAvoiddevice complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The coded video sequence is divided into multiple access units, each capable of having independent picture size parameters. This segmentation allows different parts of the video to have different resolutions based on scene activity, improving adaptability while managing complexity through modular organization.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The picture size parameters are made dynamic by allowing changes at access unit boundaries rather than being fixed for the entire sequence. The decoder can adaptively adjust resolution based on scene activity detection, enabling the system to respond to changing conditions while maintaining manageable complexity through controlled dynamic behavior.

Inventive Principle:

Principle #15Dynamics

2Productivity

If different resampling factors are applied to semantically independent pictures, then compression efficiency is improved and bandwidth is reduced, but device complexity increases

Engineering Contradiction:

Improvecompression efficiencyVSAvoiddevice complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

Different resampling factors are applied to different semantically independent pictures based on their specific scene activity characteristics. High-activity regions use lower resolution to save bandwidth, while low-activity regions maintain higher resolution, optimizing compression efficiency locally without uniformly increasing system complexity.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The picture size parameters are changed dynamically for different access units based on scene activity analysis. By modifying resolution parameters locally rather than globally, the system achieves improved compression efficiency while managing device complexity through targeted parameter adjustments rather than system-wide changes.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If reference picture resampling is performed, then adaptability to different resolutions is improved, but loss of information increases due to resampling artifacts

Engineering Contradiction:

Improveadaptability to different resolutionsVSAvoidloss of information

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

Reference pictures are resampled in advance before being used for prediction, allowing the decoder to prepare appropriate resolution versions ahead of time. This preliminary resampling enables adaptive resolution changes while managing information loss by pre-processing reference pictures with appropriate quality levels for their intended use.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Resampled reference pictures serve as intermediaries between the original high-resolution source and the current picture being decoded. By using pre-resampled reference pictures at appropriate resolutions, the system achieves adaptability to different resolutions while minimizing information loss through controlled intermediate representations rather than direct resampling of final output.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20230254506A1Method for parameter set reference constraints in coded video stream

Publication Date: 2023.08.10 TENCENT AMERICA LLC

AI summary

There is included a method and apparatus comprising computer code configured to cause a processor or processors to perform obtaining video data comprising data of a plurality of semantically independent source pictures, determining, among the video data, whether references are associated with any of a first access unit (AU) and a second AU according to at least one picture order count (POC) signal value included with the video data, and outputting a first quantity of the references set to the first AU and a second quantity of the references set to the second AU based on the at least one POC signal value.