Video Coding Picture Unit APS Placement Constraints

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In the Versatile Video Coding (VVC) standard, the current constraints on adaptive parameter sets (APS) lead to inefficiencies, such as the need to store multiple versions of APS in memory, and require additional decoding operations and bitstream rewriting during random access decoding.

Innovation Solution

The proposed method involves modifying the syntax structure by prohibiting the inclusion of prefix APS NAL units after the first NAL unit and suffix APS NAL units before the last NAL unit in a picture unit, thereby reducing the situations where two versions of an APS are required.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple versions of APS are stored in memory to handle different slices, then the coding standard supports flexible parameter configuration, but the memory requirements and device complexity increase

Engineering Contradiction:
Improveparameter configuration flexibilityVSAvoidmemory requirements
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent merges multiple APS versions into a single APS structure by allowing different slices to reference the same APS identifier. The syntax structure is modified so that all slices within a picture unit share a common APS, eliminating the need to store multiple separate APS versions in memory while maintaining the ability to configure parameters for different slices through slice_header syntax elements.

Inventive Principle:
Principle #5Merging (Combining)

2Adaptability or versatility

If multiple versions of APS are maintained in memory, then different slices can have different parameter sets, but the decoding operations and bitstream rewriting complexity increase

Engineering Contradiction:
Improveslice-specific parameter supportVSAvoiddecoding operations
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by establishing a single APS version at the beginning of the picture unit that all subsequent slices will reference. The APS is decoded once before slice decoding, and slice headers contain references to this pre-decoded APS rather than requiring separate APS decoding for each slice. This eliminates the need for bitstream rewriting operations during random access decoding.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If prefix APS NAL units are allowed after the first NAL unit, then more flexible APS placement is possible, but the situations requiring multiple APS versions increase

Engineering Contradiction:
ImproveAPS placement flexibilityVSAvoidnumber of APS versions
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies local quality by restricting prefix APS NAL units to appear only at specific locations (before the first VCL NAL unit or at defined boundaries) rather than allowing them anywhere in the bitstream. This localized constraint on APS placement prevents the creation of multiple APS versions while still providing sufficient flexibility for random access and incremental decoding scenarios.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20250056062A1Video coding and decoding
Publication Date: 2025.02.13 CANON KK
  • US20250056062A1 patent drawing
  • US20250056062A1 patent drawing
  • US20250056062A1 patent drawing

AI summary

A sequence of images is encoded in a bitstream as a series of picture units PU-01˜03. Each picture unit corresponds to one encoded image and includes one or more network abstraction layer (NAL) units NAL-01˜23. The NAL units may be video coding layer (VCL) NAL units which each contain encoded image data or adaptation parameter set NAL units which each contain an adaptation parameter set (APS) having parameters for performing one or more types of processing operation on the image data contained in one or more VCL NAL units. The APS NAL units may be prefix APS NAL units P-APS or suffix APS NAL units S-APS. An additional constraint is applied to the bitstream prohibiting inclusion of a prefix APS NAL unit after the first NAL unit of the picture unit concerned. This can avoid more than one APS applying to slices belonging to the same picture unit.