Intra-Prediction Smoothing Control for Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video coding standards face challenges in efficiently compressing screen content due to its sharp edges and textures, which can negatively impact intra-prediction compression efficiency, particularly in High Efficiency Video Coding (HEVC) and emerging standards like Versatile Video Coding (VVC).

Innovation Solution

The method involves signaling intra-prediction smoothing control parameters through a dedicated flag in a bitstream, allowing for the enabling or disabling of smoothing operations at a slice level, using a control parameter at the Sequence Parameter Set (SPS) level and another at the slice header level, to adapt intra-prediction processing to the content properties.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If intra-prediction smoothing tools (such as four-tap subpixel intra interpolation filtering and PDPC) are applied to screen content, then prediction accuracy may improve, but compression efficiency deteriorates due to sharp edges and textures

Engineering Contradiction:
Improveprediction accuracyVSAvoidcompression efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent applies dynamics by making the smoothing tool activation state changeable based on content type. A flag in the bitstream allows the decoder to dynamically enable or disable the smoothing tool for different content types (screen content vs. other content), optimizing the balance between prediction accuracy and compression efficiency for each specific case

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of smoothing tool activation by introducing a controllable flag that can be set to different values (enabled/disabled). This parameter change allows the system to adapt the smoothing behavior to match the characteristics of the input content, preventing unnecessary smoothing on screen content with sharp edges while allowing it for content types that benefit from smoothing

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If smoothing operations are applied to all content uniformly, then processing simplicity is maintained, but compression efficiency is reduced for screen content with sharp edges

Engineering Contradiction:
Improveprocessing simplicityVSAvoidcompression efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent applies local quality by differentiating the processing behavior based on content type identification. The system maintains simple uniform processing for most content while applying enhanced or disabled processing locally for screen content identified through the flag, optimizing compression efficiency for specific regions or content types without complicating the overall processing architecture

Inventive Principle:
Principle #3Local quality

3Device complexity

If no smoothing control mechanism is provided, then bitstream complexity is minimized, but adaptability to different content properties is insufficient

Engineering Contradiction:
Improvebitstream complexityVSAvoidadaptability to content properties
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent applies segmentation by dividing the control information into separate components: a flag element that indicates smoothing tool activation and separate content type identification mechanisms. This segmentation allows for flexible adaptability to different content properties while keeping the bitstream structure organized and manageable, avoiding unnecessary complexity by only including control elements when needed

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3841752B1Methods for adapting intra prediction smoothing processing for video coding, encoder, decoder and computer program product
Publication Date: 2024.09.25 HUAWEI TECH CO LTD
  • EP3841752B1 patent drawingFigure 1A
  • EP3841752B1 patent drawingFigure 1B
  • EP3841752B1 patent drawingFigure 2

AI summary

The present disclosure provides methods and devices of signaling intra predicting smoothing control parameters for encoding or decoding of video or image data. The method comprises: obtaining a first control parameter in a Sequence parameter set, SPS level, wherein a first value of the first control parameter in the SPS level indicates that smoothing is not applied for any block of a sequence, a second value of the first control parameter in the SPS level indicates that smoothing is applied for each block of the sequence, and the method further comprises: when the value of the first control parameter in the SPS level is equal to a predefined value, obtaining a second control parameter in a slice header, wherein a first value of the second control parameter in the slice header indicates that smoothing is applied for each block if the slice.