Holographic Display Aberration Correction via Neural Network Kernel
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Holographic display systems face image quality deterioration due to optical aberrations, necessitating a method to compensate for these distortions.
Innovation Solution
A method involving a neural network and kernel generation to model and optimize aberrations in holographic display devices, using a Zernike polynomial to correct aberrations across hologram plane segments, iteratively updating the kernel and neural network based on loss functions to improve image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional holographic display methods are used, then the system structure is simple, but optical aberrations cause image quality deterioration
Solution Approach 1:
The patent applies preliminary action by pre-training a neural network model to predict and compensate for optical aberrations before hologram generation. The aberration compensation process is performed in advance through kernel generation and neural network training, allowing the system to correct image quality issues before actual holographic display without adding complex real-time correction hardware
Solution Approach 2:
The patent replaces traditional mechanical or optical correction systems with a computational approach using neural networks. Instead of using complex optical components to correct aberrations, the system uses software-based neural network models that process and correct holographic data computationally, thereby improving image quality without proportionally increasing device complexity
2Measurement precision
If the entire hologram plane is processed at once, then the processing is simple, but the computational load increases and accuracy decreases
Solution Approach 1:
The patent applies segmentation by dividing the hologram plane into multiple segments and processing each segment separately through the neural network. This allows the system to model aberrations with higher precision for each local region while reducing the computational power required for each individual processing step, as each segment is handled independently rather than processing the entire high-resolution hologram plane at once
Data Source
AI summary
Provided is a method of generating a hologram, the method including generating a kernel and a neural network configured to model an aberration of a holographic display device, obtaining second image data output from the neural network to which first image data obtained by propagating a first hologram based on the kernel is input, updating the kernel and the neural network based on comparing the second image data and predetermined image data, and generating a second hologram based on the kernel and the neural network.