Intra Prediction Using Virtual Pixels for Video Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image compression methods, such as those in MPEG and H.264/AVC, face inefficiencies in encoding and decoding high-resolution video content, particularly due to the fixed size of macroblocks, which leads to increased data and reduced compression efficiency.

Innovation Solution

The proposed solution introduces a new intra prediction method that uses adjacent pixels for improved encoding efficiency by obtaining virtual pixels through linear interpolation, allowing for adaptive prediction values based on image characteristics, thereby optimizing the size and shape of coding units.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If fixed-size macroblocks are used for encoding, then the encoding process is simple, but compression efficiency deteriorates for high-resolution content

Engineering Contradiction:
Improveencoding simplicityVSAvoidcompression efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent divides the image into prediction units of various sizes (e.g., 4x4, 8x8, 16x16, 32x32 pixels) rather than using fixed macroblocks. This segmentation allows the encoder to adapt the block size to the local image characteristics, improving compression efficiency while maintaining manageable encoding complexity through hierarchical evaluation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic adaptation by selecting different prediction modes and block sizes based on image characteristics. The encoder dynamically chooses from multiple intra prediction modes (e.g., planar, angular, DC) and block sizes to optimize compression for each region, transforming the static fixed-macroblock approach into a dynamic adaptive system.

Inventive Principle:
Principle #15Dynamics

2Productivity

If more prediction modes are used to improve compression efficiency, then coding efficiency improves, but device complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprediction method complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies different prediction modes to different regions of the image based on local characteristics. For example, smooth regions use planar or DC prediction, while edges use angular predictions with specific directions. This local adaptation improves coding efficiency without requiring all complex modes to be applied uniformly across the entire image, thus managing device complexity.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent changes parameters such as prediction mode indices and block sizes based on image content analysis. By adjusting these parameters dynamically according to local image characteristics (e.g., gradient magnitude, texture complexity), the system achieves high coding efficiency while avoiding the unnecessary complexity of applying all possible modes everywhere.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If virtual pixels are obtained through linear interpolation using adjacent pixels, then prediction accuracy improves, but computational complexity increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidcomputation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs linear interpolation to obtain virtual pixels only when and where needed for specific prediction modes, rather than computing all possible virtual pixels for all blocks. This partial application of interpolation reduces computational complexity while maintaining prediction accuracy in regions where it provides benefit.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent uses virtual pixels as intermediary values obtained through linear interpolation of adjacent pixels. These virtual pixels serve as mediators that enable accurate prediction at block boundaries and corners without requiring direct computation of all pixel values, thus improving prediction accuracy while managing computational load through efficient interpolation formulas.

Inventive Principle:
Principle #24Intermediary (Mediator)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances image compression efficiency by dynamically adjusting coding units according to image characteristics, reducing the number of macroblocks and improving data compression, especially for high-resolution content.

Implementation Method 1

obtaining a first prediction value of the current predicted pixel via linear interpolation using the first virtual pixel and an adjacent left pixel on the same line as the current predicted pixel; obtaining a second prediction value of the current predicted pixel via linear interpolation using the second virtual pixel and an adjacent upper pixel on the same column as the current predicted pixel

Methodology Applied
Scientific EffectLinear interpolation:

Data Source

PatentUS10506250B2Method and apparatus for image encoding and decoding using intra prediction
Publication Date: 2019.12.10 SAMSUNG ELECTRONICS CO LTD
  • US10506250B2 patent drawing
  • US10506250B2 patent drawing
  • US10506250B2 patent drawing

AI summary

Provided are a method and apparatus for intra predicting an image, which generate a prediction value via linear interpolation in horizontal and vertical directions of a current prediction unit. The method includes: generating first and second virtual pixels by using at least one adjacent pixel located upper right and lower left to a current prediction unit; obtaining a first prediction value of a current pixel via linear interpolation using an adjacent left pixel located on the same line as the first virtual pixel and the current pixel; obtaining a second prediction value of the current pixel via linear interpolation using an adjacent upper pixel located on the same column as the second virtual pixel and the current pixel; and obtaining a prediction value of the current pixel by using the first and second prediction values.