Virtual Camera Grouping for Efficient 3D Data Generation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for generating three-dimensional data from two-dimensional image data are inefficient when dealing with large-scale viewing spaces, as they acquire and process data from all cameras, making it unnecessary to acquire data from virtual cameras that do not contribute to the free-viewpoint image generation.

Innovation Solution

An information processing system that encodes and decodes data using a global table and group table to selectively acquire and process data from virtual cameras, sorting them into groups based on internal and external parameters to efficiently generate three-dimensional data for free-viewpoint image generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If data from all cameras is acquired for generating three-dimensional data in large-scale viewing spaces, then the completeness of viewing space coverage is improved, but the data acquisition and processing load increases significantly

Engineering Contradiction:
Improveviewing space coverageVSAvoiddata processing efficiency
Core Design Contradiction:
Area of stationary objectVSProductivity

Solution Approach 1:

The patent divides the large-scale viewing space into multiple local viewing spaces, each corresponding to a group of virtual cameras. By segmenting the overall data processing task into smaller regional tasks, the system can process data for each local space independently and efficiently, rather than handling all camera data simultaneously for the entire large-scale space.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and identifies only the necessary virtual cameras that contribute to generating free-viewpoint image data for each local viewing space. By taking out and selecting only the relevant camera data needed for specific local regions, the system avoids processing unnecessary data from cameras that do not contribute to the current viewing space generation.

Inventive Principle:
Principle #2Taking out (Extraction)

2Manufacturing precision

If data from all virtual cameras is processed, then the accuracy of three-dimensional data generation is improved, but the processing time and computational resources increase

Engineering Contradiction:
Improvethree-dimensional data accuracyVSAvoiddata processing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent extracts and selects only the necessary virtual cameras that actually contribute to generating free-viewpoint image data for each local viewing space. This selective extraction maintains processing accuracy by including all relevant cameras while excluding unnecessary ones, thereby reducing processing time without sacrificing the quality of three-dimensional data generation.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by processing only the subset of camera data that is necessary for each local viewing space rather than processing all available camera data. This partial processing approach is sufficient to achieve accurate three-dimensional data generation for the target local regions while significantly reducing computational overhead and processing time.

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If all camera data is acquired for every local viewing space, then the completeness of free-viewpoint image generation is improved, but the data transmission and storage requirements increase

Engineering Contradiction:
Improvefree-viewpoint image generation completenessVSAvoiddata transmission volume
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts and identifies only the necessary virtual cameras that contribute to free-viewpoint image generation for each local viewing space. By transmitting and storing only this selected subset of camera data rather than all camera data, the system reduces data transmission volume and storage requirements while maintaining the completeness and reliability of free-viewpoint image generation for the target regions.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies local quality by optimizing data acquisition and transmission for each local viewing space independently. Each local space receives only the camera data necessary for its specific region, rather than all camera data being transmitted uniformly. This localized approach ensures reliable free-viewpoint image generation for each region while minimizing overall data transmission and storage requirements.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3410387B1Information processor and information-processing method
Publication Date: 2022.11.02 SONY GROUP CORP
  • EP3410387B1 patent drawingFigure 1
  • EP3410387B1 patent drawingFigure 2
  • EP3410387B1 patent drawingFigure 3

AI summary

The present disclosure relates to an information processing apparatus and an information processing method that are configured to be capable of efficiently acquiring information for use in generating three-dimensional data from two-dimensional image data. A grouping block sorts two or more virtual cameras for acquiring two-dimensional image data into two or more groups. A global table generation block generates a global table in which group information related with each of two or more groups is registered. A group table generation block generates, for each group, a group table in which camera information for use in generating three-dimensional data from two-dimensional image data acquired by a virtual camera sorted into a group is registered. The present disclosure is applicable to an encoding apparatus and the like, for example.