CNT Video Coding Using Activity-Based Transform Configuration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video compression techniques face challenges in achieving high-quality prediction for next-generation video content with high spatial resolution and frame rate, as they cannot effectively utilize statistical dependencies and rely solely on initial data for prediction, leading to inefficient compression and processing.

Innovation Solution

The implementation of a Conditionally Non-linear Transform (CNT) coding technique that configures CNT configuration information based on activity information, allowing for optimized prediction using all previously reconstructed signals and considering pixel correlations in the transform domain, combining prediction and transform coding for enhanced efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If prediction coding uses only initial data for prediction, then the prediction process is simple and fast, but the prediction quality is low and cannot utilize statistical dependencies

Engineering Contradiction:
Improveprediction qualityVSAvoidprediction process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by performing transform coding before prediction coding. The transform coefficients are calculated in advance and stored, then used to generate multiple prediction signals with different characteristics. This allows the prediction process to leverage pre-computed statistical information and achieve high prediction quality without requiring complex real-time analysis during the prediction step itself.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If transform coding depends only on first available data, then the coding process is efficient, but the prediction value quality is insufficient

Engineering Contradiction:
Improveprediction value qualityVSAvoidcoding efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the prediction process into multiple independent prediction signals, each generated from different transform coefficients or different combinations of transform coefficients. This segmentation allows each prediction signal to be optimized independently and enables the system to select or combine the most effective predictions, thereby improving overall prediction value quality while maintaining coding efficiency through parallel processing.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If next generation video content is processed with high spatial resolution and frame rate, then the video quality is high, but memory storage and processing requirements increase significantly

Engineering Contradiction:
Improvevideo qualityVSAvoidmemory storage requirement
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent applies parameter changes by adapting the prediction and transform processes to the specific characteristics of high-resolution video content. Different prediction modes and transform parameters are selected based on the content type, spatial frequency, and temporal characteristics. This allows the system to maintain high video quality while optimizing memory usage and processing requirements by applying more sophisticated methods only where necessary rather than uniformly across all data.

Inventive Principle:
Principle #35Parameter changes

4Productivity

If prediction coding does not utilize statistical dependency, then the coding process is simple, but compression efficiency is low

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcoding process complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent merges prediction coding and transform coding into a unified process where transform coefficients are used to generate prediction signals that are then used for residual calculation. This merging allows the system to combine the statistical dependency exploitation capabilities of transform coding with the prediction capabilities of prediction coding, achieving high compression efficiency while managing complexity through integrated processing rather than separate stages.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS10523945B2Method for encoding and decoding video signal
Publication Date: 2019.12.31 LG ELECTRONICS INC
  • US10523945B2 patent drawing
  • US10523945B2 patent drawing
  • US10523945B2 patent drawing

AI summary

The present invention provides a method for encoding a video signal comprising the steps of: obtaining activity information from the video signal, wherein the activity signal indicates information relating to edge characteristics of an image and includes edge orientation information and/or edge level information; determining conditionally non-linear transform (CNT) configuration information on the basis of the activity information; and performing CNT prediction coding on the basis of the CNT configuration information, wherein the CNT prediction coding involves performing prediction using all previously decoded pixel values.