Projected Motion Vectors for 3D Video Coding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current video coding standards like H.264, H.265/HEVC, and VP9 face inefficiencies in image data compression due to high computational loads from brute search methods for motion estimation, especially in three-dimensional video systems with multiple cameras, leading to delays and increased processing time.

Innovation Solution

The method involves projecting motion vectors from a reduced set of camera views onto other camera views using geometric projections, reducing the need for extensive motion estimation computations and searches, thereby decreasing the computational load and processing time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If brute search methods are used for motion estimation, then motion vectors can be accurately determined, but computational load increases significantly causing delays in the coding process

Engineering Contradiction:

Improvemotion vector accuracyVSAvoidcoding speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent applies preliminary action by using motion vectors from a reduced set of camera views as pre-computed reference data before encoding the full 3D video sequence. These preliminary motion vectors are then projected to generate initial motion estimates for other views, significantly reducing the search space and computational requirements for subsequent motion estimation operations.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If motion vectors are computed separately for each camera view, then each view maintains independent accuracy, but the total computational load multiplies across all camera views

Engineering Contradiction:

Improvemotion vector accuracy per viewVSAvoidtotal computational energy

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent merges the motion estimation processes across multiple camera views by selecting a reduced set of representative views for full motion vector computation. The motion information from these selected views is then combined through projection operations to generate motion vectors for all other views, thereby merging what would otherwise be separate computational tasks into a unified, more efficient process.

Inventive Principle:

Principle #5Merging (Combining)

3Productivity

If spatial techniques are used to compute motion vectors from neighbor blocks, then the number of searches is reduced, but computational delays still occur in known coding systems

Engineering Contradiction:

Improvenumber of searchesVSAvoidcoding delay

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent applies dimensionality change by transitioning from traditional 2D spatial motion estimation within a single view to leveraging motion information across the additional dimension of multiple camera views. By projecting motion vectors from one view to another using geometric relationships, the system exploits the extra spatial dimension provided by multi-view geometry to reduce computational requirements while maintaining accuracy.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10356417B2Method and system of video coding using projected motion vectors

Publication Date: 2019.07.16 INTEL CORP

AI summary

Techniques related to video coding perform by using projected motion vectors.