Asynchronous 3D Image Generation Reducing Eye Tracking Delays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D video displays, especially glasses-free types, suffer from deteriorated image quality when the user's eye positions change, due to synchronous detection and rendering processes that accumulate time delays, leading to crosstalk and image quality deterioration.
Innovation Solution
An asynchronous method and apparatus for generating 3D images that detect and render eye positions in parallel, using an eye tracker and stereo image generator to estimate current eye positions based on stored data and generate stereo images, allowing for different operation cycles based on content types and motion vectors, thereby reducing time delays and maintaining image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If synchronous detection and rendering processes are used, then system complexity is reduced, but time delays accumulate causing image quality deterioration
Solution Approach 1:
The patent divides the 3D image generation system into independent parallel processes: eye position detection, stereo image generation, and 3D image rendering. Each process operates with its own cycle independently, eliminating the accumulation of time delays that occur in synchronous systems while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The patent implements dynamic operation where the eye tracker, stereo image generator, and renderer operate at different speeds and cycles adapted to their respective processing requirements. The system dynamically selects and combines data from different time points rather than requiring synchronized operation, allowing each component to optimize its performance independently.
2Manufacturing precision
If eye position detection and rendering are performed in parallel with asynchronous cycles, then time delays are reduced improving image quality, but device complexity increases
Solution Approach 1:
The patent introduces a controller as an intermediary that manages the asynchronous parallel processes. The controller coordinates the eye tracker, stereo image generator, and renderer by selecting appropriate data from different time points and combining them, thereby managing the complexity of asynchronous operation without requiring tight synchronization between components.
Solution Approach 2:
The patent creates a universal 3D image generation system that can handle multiple content types (2D video, 3D video, still images) through the same asynchronous parallel architecture. The system universally applies the parallel processing approach across different content types, with the controller adapting the operation cycles based on content requirements.
3Ease of operation
If synchronous processes are used, then ease of operation is maintained, but productivity decreases due to accumulated time delays
Solution Approach 1:
The patent implements preliminary action by detecting eye positions continuously in advance and storing them for later use. The eye tracker operates independently ahead of time, capturing eye position data that can be retrospectively matched with rendered frames, thereby eliminating waiting time and improving processing efficiency without complicating operation.
Solution Approach 2:
The patent maintains continuity of useful action by having all three processes (eye tracking, stereo image generation, and rendering) operate continuously in parallel rather than sequentially. Each process runs independently without idle waiting time, maximizing productivity while the controller ensures proper data combination maintains operational simplicity.
Data Source
AI summary
A method for generating a three-dimensional (3D) image may detect a current eye position of a user and render a 3D image based on at least one of a previously detected eye position of the user and previously generated stereo images. A cycle at which the current eye position of user is detected and a cycle at which a 3D image is rendered may be asynchronous.