Fusion-Based Intra Transform Selection Using HoG and VIPM

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding systems face challenges in efficiently compressing digital video signals due to inefficiencies in predicting video blocks and selecting appropriate transform sets, leading to suboptimal compression and decoding performance.

Innovation Solution

The implementation of cross-component residual prediction using a fusion of prediction blocks, histogram of gradients (HoG), and virtual intra prediction mode (VIPM) to determine a transform set for video encoding and decoding, allowing for the selection of optimal transform sets based on prediction modes and fusion weights.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a fusion of multiple prediction blocks is used to predict a video block, then prediction accuracy is improved, but computational complexity increases

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

Solution Approach 1:

The patent determines HoG and VIPM from prediction blocks before the fusion operation. By performing these calculations in advance on individual prediction blocks rather than on the fused result, the computational complexity is reduced while still achieving accurate transform set selection for the final fused prediction

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent processes each prediction block separately to determine its HoG and VIPM, then fuses the prediction blocks. This segmentation approach allows the system to handle complex multi-block predictions by breaking them into manageable individual block analyses, reducing overall computational burden

Inventive Principle:
Principle #1Segmentation

2Productivity

If multiple transform sets are selected based on VIPM and HoG, then compression performance is improved, but processing time increases

Engineering Contradiction:
Improvecompression performanceVSAvoidprocessing time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent determines HoG and VIPM from prediction blocks before the fusion operation, enabling early selection of appropriate transform sets. This preliminary determination reduces processing time by avoiding repeated HoG calculations on fused blocks while still achieving optimal compression through multiple transform set selection

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses HoG parameters to determine VIPM, which then guides transform set selection. By changing the approach from direct transform selection to VIPM-based selection, the system achieves better compression performance while managing processing time through parameter-driven decision making

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If HoG is determined from the first prediction block before fusion, then transform set selection accuracy is improved, but computational overhead increases

Engineering Contradiction:
Improvetransform set selection accuracyVSAvoidcomputational overhead
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent determines HoG from the first prediction block before the fusion operation. This preliminary HoG determination provides sufficient information for accurate transform set selection without requiring HoG calculations on all fused prediction blocks, thereby reducing computational overhead while maintaining selection accuracy

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4676031A1Multiple transform set selection for fusion based intra modes
Publication Date: 2026.01.07 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP4676031A1 patent drawingFigure 1A
  • EP4676031A1 patent drawingFigure 1B
  • EP4676031A1 patent drawingFigure 1C

AI summary

Systems, devices, and methods are described herein for multiple transform set selection for fusion based intra modes. An example device may determine that a fusion of a plurality of prediction blocks is to be used to predict a video block. The device may determine a prediction block of the plurality of prediction blocks using a prediction mode. The device may determine a histogram of gradients (HoG) based on the prediction block, for example, before the fusion of the plurality of prediction blocks. The device may determine a virtual intra prediction mode (VIPM) based on the HoG. The device may determine a transform set based on the VIPM. The device may decode the video block using one of a plurality of transform sets, wherein the plurality of transform sets comprises the determined transform set.