Resampled Picture Slice Definitions for Scalable Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Video coding techniques face inefficiencies in maintaining raster scan processing order and preventing partial coding tree units when upscaling base layers to create enhancement layers, leading to coding inefficiencies and processing complexities.

Innovation Solution

The solution involves resampling video information to create a resampled picture with modified slice definitions that satisfy raster scan processing rules, including rounding partial largest coding units and dividing slices into sub-slices to ensure each slice is processed in its entirety, thereby maintaining raster scan order and preventing partial coding tree units.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If base layer is upsampled to create enhancement layer, then spatial resolution is improved, but slice definition compliance and processing order are violated

Engineering Contradiction:
Improvespatial resolutionVSAvoidslice definition compliance
Core Design Contradiction:
Manufacturing precisionVSReliability

Solution Approach 1:

The patent divides the picture into multiple slices and further divides slices into sub-slices to ensure that each slice boundary aligns with complete coding tree units. This segmentation approach allows the upsampled enhancement layer to maintain proper slice definition compliance while preserving the higher spatial resolution, resolving the contradiction between resolution improvement and slice definition reliability.

Inventive Principle:
Principle #1Segmentation

2Reliability

If slice boundaries are adjusted to prevent partial coding tree units, then processing reliability is improved, but slice structure complexity increases

Engineering Contradiction:
Improveprocessing orderVSAvoidslice structure
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces sub-slicing by dividing existing slices into multiple sub-slices, where each sub-slice contains complete coding tree units. This segmentation maintains reliable processing order without fundamentally changing the overall slice structure, as sub-slices are organizational divisions within the existing slice framework rather than a complete structural overhaul.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary adjustment of slice and sub-slice boundaries before encoding, ensuring they align with complete coding tree units. This preliminary action prevents partial coding tree units from occurring during processing, maintaining processing reliability without requiring complex runtime adjustments or structural changes during actual video encoding.

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If partial coding tree units are allowed, then encoding flexibility is improved, but decoding correctness deteriorates

Engineering Contradiction:
Improveencoding flexibilityVSAvoiddecoding correctness
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent segments the picture into slices and sub-slices with boundaries that strictly align with complete coding tree units. This segmentation provides a structured framework that maintains decoding correctness by ensuring each slice and sub-slice contains complete, independently decodable units, while still allowing flexibility in how the picture is partitioned into different slice and sub-slice configurations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the parameters of slice and sub-slice boundaries to specifically align with complete coding tree unit boundaries. This parameter adjustment ensures that encoding flexibility is maintained through different possible boundary configurations, while decoding correctness is guaranteed by the strict alignment constraint that prevents partial coding tree units.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2941882B1Inter-layer reference picture generation for hls-only scalable video coding
Publication Date: 2017.04.12 QUALCOMM INC
  • EP2941882B1 patent drawingFigure 1
  • EP2941882B1 patent drawingFigure 2A
  • EP2941882B1 patent drawingFigure 2B

AI summary

An apparatus for coding video information may include a memory unit configured to store video information associated with a picture and a processor in communication with the memory unit configured to resample video information of a reference picture to obtain a resampled picture having a plurality of slices and a different picture size than a picture to be encoded. Further, the processor may determine slice definitions for slices in the resampled picture. The slices of the resampled picture may correspond to slices of the reference picture. The processor may determine, based on the slice definitions, whether a slice of the resampled picture satisfies one or more slice definition rules. In response to determining that the slice of the resampled picture does not satisfy at least one slice definition rule, the processor can modify the slice definition for the slice so as to satisfy the slice definition rule.