Neural Network Coordinate Conversion for VR Distortion Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining image distortion parameters in VR display apparatuses require extensive data searches and data fitting, leading to slow image rendering and poor image quality due to inaccuracies in distortion information provided by manufacturers.
Innovation Solution
An optical distortion measuring apparatus that uses a neural network-based coordinate conversion model to directly calculate screen coordinates from image plane coordinates, eliminating the need for large data searches and data fitting by training the model with distortion sample information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional data search and fitting methods are used to determine image distortion parameters, then measurement precision can be achieved, but image rendering speed becomes slow
Solution Approach 1:
The patent applies preliminary action by pre-collecting distortion sample information from multiple angles of view and training the neural network model in advance. This preprocessing allows the system to bypass time-consuming data searches and fitting operations during actual image rendering, achieving both high precision distortion parameters and fast rendering speed through the pre-trained model's direct coordinate conversion.
2Device complexity
If manufacturer-provided distortion information is used, then device complexity is reduced, but measurement precision deteriorates due to inaccuracies
Solution Approach 1:
The patent implements self-service by enabling the system to automatically collect its own distortion sample information through the image collector and processing apparatus, and to autonomously train the neural network model using this self-acquired data. This eliminates dependence on potentially inaccurate manufacturer-provided distortion information while maintaining system simplicity, as the apparatus performs self-calibration without requiring external intervention or complex additional hardware.
Data Source
AI summary
An optical distortion measurement method that including receiving target field of view information; processing the target field of view information using a coordinate conversion model to obtain screen coordinates of a target angle of view, wherein the coordinate conversion model may be configured to convert image plane coordinates of the target angle of view to screen coordinates of the target angle of view; and outputting at least the screen coordinates of the target angle of view.