Eye Tracking Distortion Correction for Extended Reality Displays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional extended-reality devices, such as VR, AR, and MR devices, fail to accommodate misplacement and misalignment of a user's eyes relative to their optical axes, leading to image distortion.
Innovation Solution
A display apparatus comprising multiple projectors and tracking mechanisms to determine and correct per-pixel distortions based on the current position and orientation of each eye, applying transformations to generate distortion-corrected images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional specialized devices are calibrated according to interpupillary distance, then the device can be manufactured with fixed optical axes, but it cannot accommodate misplacement and misalignment of user's eyes relative to optical axes
Solution Approach 1:
The patent applies dynamics by making the optical system adaptable through real-time eye tracking and dynamic distortion correction. The device transitions from a static fixed-optical-axis design to a dynamic system that adjusts image rendering based on actual eye position and orientation, allowing accommodation of misplacement without changing the physical hardware structure.
Solution Approach 2:
The patent changes parameters by dynamically adjusting image distortion correction parameters based on eye tracking data. Instead of using fixed calibration parameters, the system modifies rendering parameters in real-time to compensate for varying eye positions and orientations relative to the optical axes.
2Measurement precision
If real-time eye tracking and distortion correction are implemented, then image distortion caused by eye misalignment is eliminated, but computational processing requirements increase
Solution Approach 1:
The patent replaces complex mechanical adjustment mechanisms with computational methods. Instead of physically adjusting the optical system to accommodate eye misalignment, the system uses image processing and distortion correction algorithms to achieve the same effect, substituting mechanical complexity with computational efficiency.
Solution Approach 2:
The patent implements feedback by continuously tracking eye position and orientation, using this information to dynamically adjust image rendering. The eye tracking system provides real-time feedback to the distortion correction algorithm, creating a closed-loop system that adapts to changing eye positions.
3Adaptability or versatility
If multiple displays are used to present different views to different eyes, then stereoscopic vision is achieved, but misalignment of eyes with optical axes causes distorted views
Solution Approach 1:
The patent applies local quality by providing eye-specific distortion correction for each display. Instead of using a single correction model for both eyes, the system determines and applies separate transformation parameters tailored to each eye's specific position and orientation relative to its optical axis, ensuring accurate rendering for each viewer.
Data Source
Figure 1~2
Figure 3~4B
Figure 5A~5C
AI summary
Disclosed is display apparatus (100, 200, 300) comprising first (102, 202) and second (104, 204) displays or projectors that display first and second images for first (302) and second (304) eyes, respectively; first (306) and second (308) portions facing first and second eyes, respectively, first and second portions having first (F-F') and second (S-S') optical axes, respectively; means (106, 206) for tracking positions and orientations of first and second eyes relative to corresponding optical axes, respectively; and processor (108, 208). The processor or external processor (110, 226) obtains current relative positions and orientations of both eyes; determines first and second transformations for first and second input image frames, respectively, given transformation being applied to correct apparent per-pixel distortions produced when given input image frame is displayed; and applies first and second transformations to generate first and second distortion-corrected image frames, respectively, wherein processor renders first and second distortion-corrected image frames.