Scalable Video Coding Residual Prediction Complexity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video encoding and decoding technologies face complexity and reduced coding efficiency due to residual prediction processes, particularly in spatial scalable video coding, which is challenging for real-time applications.

Innovation Solution

A method that simplifies residual prediction by splitting high-pass frames into two groups, allowing inter-layer residual prediction on frame level for one group and prohibiting it for the other, reducing encoder and decoder complexity while maintaining coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If inter-layer residual prediction is used on macroblock level, then coding efficiency is improved, but decoder complexity increases significantly

Engineering Contradiction:
Improvecoding efficiencyVSAvoiddecoder complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides high-pass frames into two groups and applies different prediction strategies to each group. Group 1 frames use inter-layer residual prediction while Group 2 frames use simple copying, thereby segmenting the complexity distribution and reducing overall decoder complexity while maintaining coding efficiency for important frames.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different quality levels of prediction to different frame groups. Group 1 frames receive full inter-layer residual prediction with higher quality, while Group 2 frames use simpler copying with lower quality, optimizing the trade-off between coding efficiency and complexity locally for each frame group.

Inventive Principle:
Principle #3Local quality

2Productivity

If adaptive residual prediction is used, then coding efficiency is improved by 5-10%, but processing complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprocessing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces a dynamic frame grouping mechanism where frames are alternately assigned to Group 1 and Group 2 based on their position in the sequence. This dynamic approach allows the system to adaptively control the balance between coding efficiency and complexity by adjusting which frames receive full prediction processing.

Inventive Principle:
Principle #15Dynamics

3Productivity

If residual prediction is enabled for every picture, then coding efficiency is maintained, but real-time decoding becomes too slow

Engineering Contradiction:
Improvecoding efficiencyVSAvoiddecoding speed
Core Design Contradiction:
ProductivityVSSpeed

Solution Approach 1:

The patent applies residual prediction only partially to Group 1 frames while using simple copying for Group 2 frames. This partial application of the complex prediction process maintains adequate coding efficiency for half the frames while dramatically reducing the overall processing load to enable real-time decoding.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP1884118B9Method and apparatus for encoding video data, and method and apparatus for decoding video data
Publication Date: 2013.06.12 THOMSON LICENSING SA
  • EP1884118B9 patent drawingFigure 1~2
  • EP1884118B9 patent drawingFigure 3~4

AI summary

Scalable video coding (SVC) based on H.264/AVC uses motion compensation, which results in motion vectors and residual pictures. To optimize the encoding of the residual pictures of the enhancement-layer, they can be predicted from the base-layer. This requires complex encoders and decoders. Simplified residual prediction techniques are provided that focus on reducing the encoding and decoding complexity of the spatial scalable enhancement-layer with optimized coding efficiency. The method for encoding video data containing high-pass frames and low-pass frames comprises the steps of encoding the low-pass frames, wherein residual prediction (ORP) may be used, splitting the high-pass frames into two interleaving frame groups, encoding the frames of the first of said frame groups, wherein residual prediction may be used, and encoding the frames of the second of the frame groups using an encoding method without residual prediction (NoRP).