Omni-Directional Image Projection for New Viewpoint Synthesis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods fail to directly acquire omni-directional images with centers different from those of previously captured images, hindering realistic virtual reality experiences.

Innovation Solution

A method and apparatus for generating an omni-directional image with a different center by aligning and projecting two-dimensional images from multiple omni-directional images, adjusting north and south poles to epipoles, and calculating disparity values to create a new center based on quantized positions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If omni-directional images are acquired from fixed camera positions, then image acquisition is simplified, but the ability to provide omni-directional images with different centers is limited

Engineering Contradiction:
Improveability to provide omni-directional images with different centersVSAvoidimage processing complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by pre-processing omni-directional images into 2D images with aligned north and south poles, and pre-calculating disparity values between different centers. This preparation work is done in advance so that when a new center is needed, the system can quickly generate the required omni-directional image without performing complex real-time processing, thus resolving the contradiction between versatility and complexity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces 2D images as an intermediary representation between the original omni-directional images and the target omni-directional images with different centers. By converting omnidirectional images to 2D format, aligning their poles, and using them as intermediate data, the system can efficiently synthesize new views without directly manipulating complex spherical coordinates, thereby reducing processing complexity while maintaining adaptability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If complex image processing is performed to generate omni-directional images with new centers, then image quality is improved, but system load and power consumption increase

Engineering Contradiction:
Improveimage qualityVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent performs computationally intensive operations in advance: converting omnidirectional images to 2D format, aligning north and south poles, and calculating disparity values between different centers. This preliminary processing ensures high image quality is achieved while the actual runtime operation requires minimal computation, thereby reducing power consumption during VR/AR device operation while maintaining reliable image quality.

Inventive Principle:
Principle #10Preliminary action

3Speed

If multiple omni-directional images are processed in real-time, then user movement response is improved, but processing time and system load increase

Engineering Contradiction:
Improveresponse to user movementVSAvoidprocessing time
Core Design Contradiction:
SpeedVSLoss of time

Solution Approach 1:

The patent performs all complex processing operations before runtime: converting multiple omnidirectional images to 2D format, aligning their poles, and pre-calculating disparity values. When the user moves, the system only needs to perform simple image selection and composition operations, achieving fast response to user movement while minimizing real-time processing time and system load.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3590090B1Method and apparatus for processing omni-directional image
Publication Date: 2026.02.25 SAMSUNG ELECTRONICS CO LTD
  • EP3590090B1 patent drawingFigure 1~2
  • EP3590090B1 patent drawingFigure 3~4
  • EP3590090B1 patent drawingFigure 5~7

AI summary

A method for processing information about an omni-directional image includes generating a first 2D image projected from a first omni-directional image, by setting points on the first omni-directional image, which intersect a straight line passing through a first position that is a center of the first omni-directional image and a second position that is a center of a second omni-directional image, to a first pole and a second pole, generating a second 2D image projected from the second omni-directional image, by setting points on the second omni-directional image, which intersect the straight line passing through the first position and the second position, to a third pole and a fourth pole, generating a third 2D image corresponding to a 2D image projected from a third omni-directional image centered in a third position between the first position and the second position, based on the first 2D image and the second 2D image.