Gaze-Based Image Rendering for VR Bandwidth Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image rendering methods for virtual reality struggle with real-time performance and computational efficiency due to high demands for display definition and data transmission, especially in virtual reality (VR) and augmented reality (AR), where existing compression methods fail to meet the requirements of high resolution and high refresh rates.
Innovation Solution
An image rendering method that determines a gaze point on a VR device screen, adjusts sampling areas based on this point, and performs resolution sampling to create high and low-resolution areas, which are then spliced and transmitted, optimizing bandwidth usage and reducing computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If compression processing is performed on the entire image, then bandwidth consumption is reduced, but real-time performance deteriorates due to large computational resource consumption
Solution Approach 1:
The image is divided into a first image area and a second image area based on the gaze point position. The first image area (peripheral region) undergoes compression processing while the second image area (central region) maintains original resolution. This segmentation allows selective compression that reduces overall bandwidth consumption without compromising the real-time performance for the critical central viewing area.
Solution Approach 2:
Different quality levels are applied to different regions of the image based on human visual characteristics. The central image area corresponding to the gaze point maintains high quality (original resolution) while the peripheral areas use compressed quality. This local quality differentiation reduces total bandwidth consumption while preserving visual quality where it matters most for real-time performance.
2Manufacturing precision
If high definition resolution is used for the entire image, then display quality is improved, but computational resource consumption increases
Solution Approach 1:
The display area is segmented into a first image area and a second image area, where the second image area (central region around gaze point) uses high definition resolution while the first image area (peripheral region) uses lower resolution. This segmentation reduces total computational resource consumption while maintaining high display quality in the critical central viewing area.
Solution Approach 2:
High definition quality is applied locally to the central image area corresponding to the gaze point, while peripheral areas use reduced quality. This local quality approach maintains manufacturing precision where it is most needed (central vision) while significantly reducing overall computational resource consumption for rendering and transmission.
3Loss of information
If the entire image is transmitted at original resolution, then image quality is maintained, but transmission data amount increases
Solution Approach 1:
The image transmission is segmented into two parts: the first image area (peripheral) is transmitted in compressed format while the second image area (central) is transmitted at original resolution. This segmentation reduces the total transmission data amount while preserving image quality in the critical central region where the user's gaze is focused.
Solution Approach 2:
Original quality transmission is applied locally to the central image area corresponding to the gaze point, while peripheral areas are transmitted with compression. This local quality approach minimizes transmission data amount by reducing redundancy in peripheral regions while maintaining image quality where it is most important for the user's visual experience.
Data Source
AI summary
An image rendering method, device, and system, a storage medium, and an image display method are provided. The image rendering method includes: acquiring an image to be displayed; according to a gaze point of human eyes on a display screen, obtaining a gaze point position on the image to be displayed; determining, according to the gaze point position, a first sampling area and a second sampling area of the image to be displayed; performing first resolution sampling on the first sampling area to obtain a first display area; performing second resolution sampling on the image to be displayed to obtain a second display area corresponding to the second sampling area, a resolution of the second sampling area being greater than that of the second display area; and splicing the first display area and the second display area to obtain an output image to be transmitted to the virtual reality device.