Video Decoding Inverse Scanning Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video data compression methods, such as H.264/MPEG-4 AVC, face inefficiencies in coding motion information and increased computational complexity due to varying block sizes and motion vector correlations, leading to higher coding bits for residual blocks.

Innovation Solution

An apparatus that decodes video data by using an entropy decoding unit, inverse scanning unit, inverse quantization unit, inverse transform unit, and inter prediction unit to generate prediction and residual blocks, with transform units larger than 4x4 blocks, and applies a diagonal scan pattern to reduce coding bits and computational complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If various sizes are used for inter prediction, then motion estimation accuracy is improved, but coding bits of residual block increase

Engineering Contradiction:
Improvemotion estimation accuracyVSAvoidcoding bits of residual block
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent divides the transform unit into multiple subsets for separate inverse scanning. This segmentation allows the residual block to be processed in smaller units, reducing the overall coding bits required while maintaining accurate motion estimation through the use of various block sizes in the prediction stage

Inventive Principle:
Principle #1Segmentation

2Productivity

If motion information of neighboring blocks is used for prediction, then coding efficiency is improved, but computational complexity increases

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

Solution Approach 1:

The patent performs preliminary inverse scanning of quantized coefficient components into subsets before the final inverse scanning to generate the quantized block. This preliminary action organizes the data structure in advance, enabling efficient motion information prediction from neighboring blocks while reducing the computational complexity of the overall decoding process

Inventive Principle:
Principle #10Preliminary action

3Productivity

If transform unit size is increased, then coding efficiency is improved, but scanning complexity increases

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

Solution Approach 1:

The patent segments the inverse scanning process into two stages: first scanning quantized coefficient components into subsets, then scanning the subsets to generate the final quantized block. This segmentation reduces scanning complexity while maintaining the benefits of larger transform unit sizes for improved coding efficiency

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3917144A1Method of decoding video data
Publication Date: 2021.12.01 GENSQUARE LLC
  • EP3917144A1 patent drawingFigure 1
  • EP3917144A1 patent drawingFigure 2
  • EP3917144A1 patent drawingFigure 3

AI summary

Provided is an apparatus that includes an inverse scanning unit configured to an inverse scan pattern to the quantized coefficient components to generate a quantized block having a size of a transform unit, an inverse quantization unit configured to generate a quantization parameter and to inverse-quantize the quantized block to generate a transformed block, an inter prediction unit configured to derive motion information and to generate a prediction block, and an adder configured to generate a reconstructed block using the residual block and the prediction block. The transform unit is larger than a 4x4 block, each of the quantized coefficient components are inversely scanned in the unit of subset to generate multiple subsets, and the multiple subsets are inversely scanned to generate the quantized block. Accordingly, the coding efficiency of the motion information is improved by including various motion vector candidates. Also, the amount of coding bits of the residual block is reduced by applying diagonal scan pattern to each subset.