Inter-layer Prediction for Scalable Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face challenges in maintaining image quality when reducing bitrate, particularly in scalable video coding, where interlayer prediction methods fail to effectively handle interlaced and progressive video signals with different scanning modes and resolutions.

Innovation Solution

The method involves setting motion-related information between intra-mode and inter-mode macroblocks, deriving motion information from adjacent macroblocks, and upsampling pictures to match resolutions for interlayer prediction, while also considering the types of pictures and macroblocks in different layers for accurate texture and motion prediction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If interlayer prediction is performed using coded information of the base layer for the enhanced layer, then coding efficiency is improved, but image quality is significantly reduced when bitrate is lowered

Engineering Contradiction:
Improvecoding efficiencyVSAvoidimage quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The video signal is divided into multiple layers (base layer and enhanced layer) with different resolutions and scanning modes. The base layer operates at lower resolution while the enhanced layer operates at higher resolution, allowing the system to segment the coding task to maintain overall image quality while achieving lower average bitrate through the base layer

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension of scalability by supporting both interlaced and progressive scanning modes across different layers. This allows the enhanced layer to use progressive scanning for higher quality while the base layer uses interlaced scanning for lower bitrate, effectively adding a quality/bitrate dimension to the coding system

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If pictures of different resolutions are used in base and enhanced layers, then adaptability to different scanning modes is improved, but interpolation complexity increases

Engineering Contradiction:
Improvescanning mode compatibilityVSAvoidinterpolation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent performs resolution adjustment (upsampling or downsampling) of the base layer picture before interlayer prediction is executed. This preliminary action prepares the base layer picture to match the resolution of the enhanced layer picture, avoiding the need for complex real-time interpolation during the prediction process and reducing overall system complexity

Inventive Principle:
Principle #10Preliminary action

3Adaptability or versatility

If interlaced video is coded into even and odd fields, then compatibility with traditional formats is improved, but motion prediction accuracy deteriorates

Engineering Contradiction:
Improveformat compatibilityVSAvoidmotion prediction accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

Instead of treating interlaced fields as separate entities that require complex field-to-field prediction, the patent inverts the approach by allowing progressive prediction from the base layer to the enhanced layer. This means using progressive scanning logic even when the base layer is interlaced, thereby simplifying motion prediction while maintaining format compatibility through proper field handling

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentEP1977608B1Inter-layer prediction method for video signal
Publication Date: 2020.01.01 LG ELECTRONICS INC
  • EP1977608B1 patent drawingFigure 1A~2A
  • EP1977608B1 patent drawingFigure 2B~2C
  • EP1977608B1 patent drawingFigure 2D

AI summary

The present invention relates to a method for conducting interlayer prediction in encoding or decoding of video signal . The present method checks whether at least one of vertically-adjacent two macro blocks on a base layer is intra mode, and if only one of the two macro blocks is identified to intra mode, motion information of an inter-mode macro block of the two macro blocks is copied to the other intra-mode macro block in order to set the intra-mode macro block to a macro block having motion information. In addition, if the two macro blocks are intra mode and contain field components respectively, two frame macro blocks are constructed from the two field macro blocks.