Derived Disparity Vector Storage for 3D Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing techniques for deriving disparity vectors in multi-view video coding face challenges such as increased storage requirements and memory accesses, particularly when using implicit disparity vectors in neighboring block-based disparity vector derivation processes.

Innovation Solution

The proposed solution involves storing derived disparity vectors (DDVs) for a slice and using them to predict disparity vectors for subsequent blocks, updating the DDVs based on coded blocks, and prioritizing DDVs from adjacent blocks, thereby reducing the need for explicit storage of implicit disparity vectors and optimizing memory access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If implicit disparity vectors are stored for each block in multi-view video coding, then disparity vector prediction accuracy is improved, but storage requirements and memory accesses increase

Engineering Contradiction:

Improvedisparity vector prediction accuracyVSAvoidstorage requirements

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the storage of disparity vectors by organizing them at the slice level rather than the block level. Derived disparity vectors (DDVs) are stored for each slice, and blocks within a slice inherit the DDV from their neighboring blocks, eliminating the need to store explicit disparity vectors for every individual block while maintaining prediction accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent merges the disparity vector storage function into the slice-level DDV structure. By combining the disparity vector information across multiple blocks within a slice and using neighboring block DDVs for prediction, the system reduces redundant storage while preserving the accuracy benefits of block-level disparity information.

Inventive Principle:

Principle #5Merging (Combining)

2Measurement precision

If explicit disparity vectors are stored for each block, then disparity vector prediction accuracy is improved, but memory accesses and processing complexity increase

Engineering Contradiction:

Improvedisparity vector prediction accuracyVSAvoidmemory access complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary action by pre-calculating and storing derived disparity vectors at the slice level before block processing. This allows subsequent blocks to access pre-computed DDV data from neighboring slices, reducing the need for complex real-time calculations and memory accesses during block-level processing.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces slice-level derived disparity vectors as an intermediary structure between the reference disparity data and individual block processing. This intermediary layer provides efficient access to disparity information without requiring direct block-level storage, simplifying memory access patterns and reducing processing complexity.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentEP2965522B1Derived disparity vector in 3D video coding

Publication Date: 2017.05.10 QUALCOMM INC

Figure 1
Figure 2
Figure 3

AI summary

A video coder stores only one derived disparity vector (DDV) for a slice of a current picture of the video data. The video coder uses the DDV for the slice in a Neighboring Block Based Disparity Vector (NBDV) derivation process to determine a disparity vector for a particular block. Furthermore, the video coder stores, as the DDV for the slice, the disparity vector for the particular block.