Light Field Surface Editing via Occluded Pixel Projection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current light field editing techniques are cumbersome and time-consuming due to the need to navigate through multiple views to select and edit surfaces, especially when dealing with occluded areas, as they are designed for 2D image editing and struggle with the multidimensional nature of light fields.

Innovation Solution

A method and apparatus that allow users to select and edit surfaces in a light field using a single reference view, projecting occluded pixels into this view for consistent editing without navigating through all views, and enabling common 2D image editing tools to be applied, such as changing color, texture, or illumination.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If users navigate through multiple views to select and edit surfaces in light field, then editing completeness is improved, but operation complexity and time consumption increase

Engineering Contradiction:
Improveediting completenessVSAvoidoperation complexity
Core Design Contradiction:
Manufacturing precisionVSEase of operation

Solution Approach 1:

The patent projects occluded pixels from multiple views into a single reference view by synthesizing information across different spatial dimensions. This transforms the multidimensional navigation problem into a two-dimensional editing task on the reference view, where occluded regions are automatically reconstructed from other views' data, eliminating the need for users to manually navigate through multiple views while maintaining editing completeness

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The system creates a synthesized copy of occluded pixel information by projecting data from multiple views into the reference view. This virtual copying of pixel values from different perspectives allows users to edit surfaces as if all views were visible simultaneously, achieving complete surface editing without actual multi-view navigation

Inventive Principle:
Principle #26Copying

2Manufacturing precision

If users navigate through multiple views to ensure consistency in light field editing, then editing accuracy is improved, but time consumption increases

Engineering Contradiction:
Improveediting accuracyVSAvoidtime consumption
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary projection of occluded pixels from multiple views into the reference view before the user begins editing. This pre-computation of pixel correspondence and value projection ensures that all necessary information is already integrated into the reference view, allowing users to edit with full accuracy knowledge without subsequent time-consuming consistency checks across multiple views

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The reference view serves as an intermediary that integrates information from multiple views. By projecting occluded pixel data through this intermediary representation, the system automatically ensures consistency across all views during editing, eliminating the need for users to manually verify consistency across multiple views while maintaining editing accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

3Ease of operation

If light field editing tools are designed for 2D images, then ease of use is improved, but adaptability to light field data deteriorates

Engineering Contradiction:
Improveease of useVSAvoidadaptability to light field data
Core Design Contradiction:
Ease of operationVSAdaptability or versatility

Solution Approach 1:

The system creates a two-dimensional projected representation of occluded light field data in the reference view, which can be manipulated using standard 2D image editing tools. This virtual copy preserves the multidimensional light field information while presenting it in a 2D format that existing tools can handle, bridging the gap between simple tool interfaces and complex light field data structures

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system transforms light field data parameters by projecting three-dimensional occluded pixel information into two-dimensional reference view coordinates. This parameter transformation allows conventional 2D editing tools to operate on light field data by changing the spatial representation parameters while maintaining the underlying light field semantics through the projection relationship

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3477596B1Method and apparatus for selecting a surface in a light field, and corresponding computer program product
Publication Date: 2021.06.09 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP3477596B1 patent drawingFigure 1(a)~1(b)
  • EP3477596B1 patent drawingFigure 2~5
  • EP3477596B1 patent drawingFigure 3A~3C

AI summary

A method for selecting at least one surface in a light field is disclosed. Such a light field is associated with multiple views corresponding to different view points of a same scene. The method comprises: - displaying (S402) one of the multiple views as a reference view; - selecting (S403) at least one surface in the reference view; - identifying (S404) a set of pixels belonging to the selected surface, which are occluded in the reference view but visible in at least one of the multiple views; - projecting (S405) the set of pixels into the reference view, on the basis of a value of the pixels of the set in the view or views in which they are visible, in order to form an upgraded reference view; - displaying (S406) the upgraded reference view; - updating (S407) said selected surface in said upgraded reference view.