Intraframe Prediction Using Lower Right Reference Pixels

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current intraframe prediction methods are limited by the inability to use pixel points on the right and below the current block as reference points due to the raster scan or Z-scan processing sequence, reducing the information available for intraframe prediction and hindering accuracy.

Innovation Solution

A novel intraframe prediction method that adjusts the coding and decoding sequence to utilize reconstructed pixel values on the lower right as reference points, allowing for two-pass coding and decoding, and employs a prediction direction-based adaptive linear weighting scheme to enhance prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional intraframe prediction uses only left and above reference pixel points, then the coding sequence can follow simple raster scan or Z-scan, but the prediction accuracy is reduced due to limited reference information

Engineering Contradiction:
Improveprediction accuracyVSAvoidcoding sequence complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by decoding all interframe predicted blocks before intraframe predicted blocks, ensuring that lower right reference pixel points are reconstructed and available before they are needed for intraframe prediction. This reordering of decoding operations allows the use of additional reference information without requiring complex real-time computations during the prediction process.

Inventive Principle:
Principle #10Preliminary action

2Loss of information

If the coding sequence is adjusted to decode interframe blocks before intraframe blocks, then lower right reference pixel points become available for intraframe prediction, but the decoding process complexity increases

Engineering Contradiction:
Improvereference information availabilityVSAvoiddecoding process complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the decoding process into two distinct phases: first decoding all interframe predicted blocks, then decoding intraframe predicted blocks. This segmentation allows the system to manage complexity by handling different prediction types in separate stages, with each stage having access to appropriate reference information without requiring complex interleaved processing.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If two-pass coding is implemented for intraframe predicted blocks, then more reference pixel points including lower right can be utilized, but the coding time and processing duration increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidcoding time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The two-pass coding approach applies preliminary action by performing a first pass to decode interframe blocks and prepare reference information, then a second pass to decode intraframe blocks using the prepared references. This preliminary preparation eliminates the need for complex real-time reference generation during intraframe decoding, reducing overall processing time despite the additional pass.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10750200B2Encoding method, decoding method, encoder, and decoder
Publication Date: 2020.08.18 PEKING UNIV SHENZHEN GRADUATE SCHOOL
  • US10750200B2 patent drawing
  • US10750200B2 patent drawing
  • US10750200B2 patent drawing

AI summary

Disclosed is an intraframe prediction method, wherein for an intraframe predicted block in an interframe predicted frame, when an interframe predicted blocks exists on the lower right of the intraframe predicted block, a coding and decoding sequence of intraframe and interframe blocks is adjusted, and intraframe prediction is performed in a lower right intraframe prediction approach, i.e., performing prediction using the lower right reconstructed pixel value as the reference pixel point for the intraframe predicted block; at the coding procedure, each coded unit is first coded, and upon completion of the coding, information of all interframe predicted blocks is written into a code stream; next, two-pass coding is performed to the intraframe predicted block; lower right intraframe coding is performed during the two-pass coding process; after completion of the two-pass coding, the information of the intraframe predicted block is written into the code stream; at the decoding procedure, decoding of all interframe predicted blocks is first completed; then, lower right intraframe decoding is performed to the intraframe predicted blocks. The present disclosure may improve the intraframe prediction accuracy and enhance the coding efficiency of the intraframe predicted block in the interframe predicted frame.