Wedgelet Partition Coding With Neighbor Line Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing coding schemes for sample arrays, including depth maps, face inefficiencies due to the need for significant side information to adapt block partitioning shapes to the content, particularly when using wedgelet-type partitioning.

Innovation Solution

Predict the position of wedgelet separation lines within a block based on neighboring blocks, refine the position using refinement information, and encode or decode the block in units of wedgelet partitions to reduce side information overhead.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If wedgelet-type partitioning is used to adapt block shapes to sample array content, then coding accuracy is improved, but side information overhead increases significantly

Engineering Contradiction:
Improvecoding accuracyVSAvoidside information overhead
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies preliminary action by predicting the wedgelet separation line position in the current block based on the already decoded wedgelet separation line position in the neighboring block. This prediction is performed before actual decoding, allowing the decoder to reconstruct the partition structure without transmitting all position information explicitly. The refinement information is then applied to adjust the predicted position, achieving accurate reconstruction with reduced overhead.

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If rectangular or quadratic blocks are used for coding, then device complexity is reduced, but adaptability to sample array content deteriorates

Engineering Contradiction:
Improvecoding structure simplicityVSAvoidblock shape adaptability
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent applies segmentation by dividing each block into two wedgelet partitions using a separation line. This allows the coding structure to adapt to content variations (improving adaptability) while maintaining a relatively simple decoding process based on linear separation (controlling complexity). The wedgelet partitioning provides flexible shape adaptation without requiring complex irregular block structures.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If wedgelet separation line positions are explicitly signaled for each block, then partitioning precision is improved, but data transmission rate increases

Engineering Contradiction:
Improveseparation line position precisionVSAvoiddata transmission rate
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies feedback by using the wedgelet separation line position from the neighboring block (already decoded) as a reference to predict the position in the current block. This creates a feedback loop where previously decoded information is utilized to reduce the information needed for current decoding. The refinement information serves as a correction to the prediction, ensuring precision while minimizing transmitted data.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS12501030B2Effective wedgelet partition coding
Publication Date: 2025.12.16 DOLBY VIDEO COMPRESSION LLC
  • US12501030B2 patent drawing
  • US12501030B2 patent drawing
  • US12501030B2 patent drawing

AI summary

Wedgelet separation lines of neighboring blocks are predicted from each other by aligning the wedgelet separation lines of such neighboring blocks to each other, thereby reducing the side information coding rate that may be used.