Omnidirectional Image Encoding for Immediate Field-of-View Display

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing systems face challenges in immediately displaying the field-of-view range of a user due to the inefficiencies in encoding and decoding omnidirectional images, particularly when the line-of-sight vector changes abruptly, as they rely on general encoding schemes like MPEG 2 and AVC that restrict random access points and hinder efficient decoding of specific image areas.

Innovation Solution

The system encodes omnidirectional images by combining high-resolution and low-resolution images across opposite surfaces of a 3D model, allowing for efficient decoding and display of only the necessary image areas corresponding to the user's line-of-sight vector, reducing the decoding load and enabling immediate image updates.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If general encoding schemes like MPEG 2 or AVC are used to encode omnidirectional images, then compression efficiency is improved, but the ability to decode only specific image areas corresponding to the user's line-of-sight vector is hindered

Engineering Contradiction:
Improvecompression efficiencyVSAvoidability to decode specific image areas
Core Design Contradiction:
Loss of energyVSEase of operation

Solution Approach 1:

The omnidirectional image is divided into multiple regions based on the user's line-of-sight vector and field-of-view range. Only the relevant regions (front, side, or back regions) are decoded and displayed, while other regions remain encoded. This segmentation allows the system to maintain high compression efficiency for the entire omnidirectional image while enabling efficient decoding of only the necessary portions.

Inventive Principle:
Principle #1Segmentation

2Speed

If the entire omnidirectional image is decoded to enable immediate display of the field-of-view range, then display responsiveness is improved, but the decoding process amount increases

Engineering Contradiction:
Improvedisplay responsivenessVSAvoiddecoding process amount
Core Design Contradiction:
SpeedVSPower

Solution Approach 1:

Instead of decoding the entire omnidirectional image, the system performs partial decoding only on the regions corresponding to the user's field of view. This partial action approach maintains display responsiveness by avoiding the need to decode unnecessary regions, thereby reducing the overall decoding process amount and computational power required.

Inventive Principle:
Principle #16Partial or excessive action

3Adaptability or versatility

If the line-of-sight vector is decided by the user, then adaptability is improved, but the recording device's ability to encode only the corresponding area is reduced

Engineering Contradiction:
Improveuser control over line-of-sightVSAvoidencoding flexibility
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system pre-encodes the entire omnidirectional image using general encoding schemes, creating a complete encoded stream that contains all possible viewing regions. When the user changes their line-of-sight vector, the reproducing device can immediately switch to decoding only the relevant pre-encoded regions without requiring the recording device to re-encode. This preliminary encoding action maintains user adaptability while simplifying the encoding process.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3267687B1Image processing device and image processing method
Publication Date: 2021.10.20 SONY GROUP CORP
  • EP3267687B1 patent drawingFigure 1~2
  • EP3267687B1 patent drawingFigure 3
  • EP3267687B1 patent drawingFigure 4~5

AI summary

[Object] To enable an image of a field-of-view range of a user to be displayed immediately. [Solution] An encoder encodes, for a pair of opposite surfaces among a plurality of surfaces of a model for an omnidirectional image, a first composite image obtained by combining an omnidirectional image of a first surface of a high resolution and an omnidirectional image of a second surface of a resolution lower than the high resolution and a second composite image obtained by combining an omnidirectional image of the second surface of the high resolution and an omnidirectional image of the first surface of a low resolution. For example, the present disclosure can be applied to an image display system or the like.