Tile-Aligned Inter-Layer Prediction for Parallel Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face challenges in parallel processing across layers, particularly in HEVC extensions, where inter-dependencies between layers increase codec delay and up-sampling or inter-layer filtering across tile boundaries are not guaranteed, leading to inefficiencies in decoding and compression.

Innovation Solution

The implementation of techniques that restrict inter-layer prediction and signaling methods to align tile boundaries across layers, allowing for efficient parallel processing by ensuring the availability of reconstructed samples and enabling up-sampling or filtering without breaking parallel processing architectures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of substance

If inter-layer prediction and filtering are performed across tile boundaries in multi-layer video coding, then compression efficiency is improved, but parallel processing capability deteriorates and codec delay increases

Engineering Contradiction:
Improvecompression efficiencyVSAvoidparallel processing capability
Core Design Contradiction:
Loss of substanceVSProductivity

Solution Approach 1:

The video picture is divided into multiple tiles, and each tile can be processed independently in parallel. By restricting inter-layer prediction and filtering to within tile boundaries or aligning tile boundaries across layers, the patent enables parallel processing of multiple tiles while maintaining compression efficiency through inter-layer prediction within each tile region.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs up-sampling and filtering operations on reference layer pictures before decoding the current layer picture. This preliminary action ensures that reconstructed samples are available when needed for inter-layer prediction, eliminating the need to wait for complete reference layer decoding and enabling parallel processing.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If complete decoding of reference layer pictures is required before performing inter-layer prediction, then prediction accuracy is improved, but codec delay increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidcodec delay
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs up-sampling and filtering of reference layer pictures in advance, before the complete decoding of the current layer picture is required. This preliminary preparation of reconstructed samples allows inter-layer prediction to proceed without waiting for full reference layer decoding, reducing codec delay while maintaining prediction accuracy through the use of pre-processed reference samples.

Inventive Principle:
Principle #10Preliminary action

3Loss of substance

If up-sampling or inter-layer filtering is performed across tile boundaries, then compression efficiency is improved, but reliability of parallel processing deteriorates

Engineering Contradiction:
Improvecompression efficiencyVSAvoidparallel processing reliability
Core Design Contradiction:
Loss of substanceVSReliability

Solution Approach 1:

The patent divides the picture into tiles and restricts up-sampling and inter-layer filtering operations to within tile boundaries or aligns tile boundaries across layers. This segmentation approach ensures that each tile can be processed independently and reliably in parallel, while still achieving compression efficiency through inter-layer prediction within the confined tile regions.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2965514B1Parallel processing for video coding
Publication Date: 2020.10.07 QUALCOMM INC
  • EP2965514B1 patent drawingFigure 1
  • EP2965514B1 patent drawingFigure 2
  • EP2965514B1 patent drawingFigure 3

AI summary

In one example, a device for coding video data includes a video coder configured to code a value for a syntax element representative of whether any two reference layer samples, collocated with two respective enhancement layer picture samples within a common enhancement layer tile, must be within a common reference layer tile, and code the enhancement layer picture samples based at least in part on the value of the syntax element.