Neural Network Interpolation for Digital Holographic Portraits

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for recording digital holographic portrait photographs are limited by the need for expensive equipment and result in unnatural, lifeless images due to restricted parallax and color accuracy, requiring lengthy and difficult-to-maintain stillness from subjects.

Innovation Solution

A method using a combination of CHIMERA or stereogram technology with neural network image interpolation, allowing for the recording of immersive, full-color holograms with a reduced number of cameras by determining the maximum allowable angle between cameras through preliminary tests and interpolating images to achieve half or full parallax.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple cameras are arranged at large angles to capture full parallax, then the parallax effect is improved, but the image quality and color accuracy deteriorate due to interpolation errors

Engineering Contradiction:

Improveparallax effectVSAvoidimage quality

Core Design Contradiction:

Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent performs preliminary tests to determine the maximum allowable angle between cameras before actual shooting. This preliminary action establishes optimal camera arrangement parameters (angles within 20-45 degrees) that balance parallax effect with image quality, preventing excessive interpolation errors while still achieving full parallax capability.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent changes the parameter of camera arrangement angles to optimize the balance between parallax and image quality. By controlling angles to be within 20-45 degrees and using neural network interpolation algorithms, the system achieves full parallax while maintaining acceptable image quality through parameter optimization.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If a plurality of cameras are used to capture images from multiple angles, then the parallax and realism are improved, but the device complexity and cost increase

Engineering Contradiction:

ImproverealismVSAvoidnumber of cameras

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent uses neural network image interpolation to generate intermediate images that copy and reconstruct the visual information between actual camera captures. This allows the system to simulate multiple camera viewpoints using fewer physical cameras, reducing device complexity while maintaining realism through intelligent image synthesis.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical solution of using many physical cameras arranged at various angles with an intelligent software-based neural network interpolation system. This substitution reduces hardware complexity while achieving the same effect of capturing multi-angle perspectives for realistic holographic portraits.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Device complexity

If cameras are arranged at maximum angles to reduce the number of cameras, then the device complexity is reduced, but the color accuracy and image fidelity worsen

Engineering Contradiction:

Improvenumber of camerasVSAvoidcolor accuracy

Core Design Contradiction:

Device complexityVSManufacturing precision

Solution Approach 1:

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent uses feedback from preliminary testing and image quality assessment to optimize camera arrangement parameters. By measuring color accuracy and image fidelity at different camera angles, the system adjusts the maximum allowable angle parameter to achieve the optimal balance between reducing camera count and maintaining color accuracy.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20240361728A1Method and apparatus for recording digital holographic portrait photograph using neural network image interpolation

Publication Date: 2024.10.31 KWANGWOON UNIVERSITY INDUSTRY ACADEMIC COLLABORATION FOUNDATION

AI summary

Provided is a method of recording a digital holographic portrait photograph, which is performed by an apparatus for recording a digital holographic portrait photograph, the method including the steps of: (a) performing a preliminary test using a plurality of cameras arranged at a specific rotation angle; (b) determining a maximum allowable angle between the cameras through the preliminary test; (c) acquiring a plurality of images using the plurality of cameras arranged at the maximum allowable angle; and (d) interpolating the plurality of images to record a holographic portrait photograph.