VR Super-Resolution via Unperceived Noise Injection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Virtual reality (VR) systems using head-mounted displays (HMDs) face challenges in achieving high resolution and visual acuity due to limitations in computational power of mobile graphics processing units (GPUs), which struggle to manage frame updates for a large number of pixels, especially in providing an immersive experience with a 180-degree field of view.
Innovation Solution
Introducing unperceived noise signals, such as zero-mean Gaussian white noise, to video frames for each eye's field of view, which are combined with the video signals to enhance dynamic range and improve image sharpness through binocular summation, allowing for a simulated resolution increase without requiring higher computational power.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the number of pixels in the display is increased to achieve higher resolution, then visual acuity is improved, but the computational power required to manage frame updates becomes excessive for mobile GPUs
Solution Approach 1:
The patent uses a lower-resolution display and generates a high-resolution virtual copy through super-resolution processing. The controlling circuitry creates virtual pixel data that mimics what a high-resolution display would produce, allowing the system to achieve high visual acuity without requiring a physically high-resolution display or excessive computational power for rendering.
Solution Approach 2:
The patent introduces an intermediary processing layer (super-resolution processing) between the low-resolution display output and the viewer's perception. This intermediary generates additional high-frequency pixel data that bridges the gap between the actual display resolution and the perceived resolution, resolving the contradiction between display hardware limitations and visual quality requirements.
2Manufacturing precision
If noise signals are added to video frames to enhance dynamic range, then image sharpness is improved, but the complexity of signal processing increases
Solution Approach 1:
The patent converts the typically harmful effect of noise into a beneficial tool for enhancement. By adding controlled noise signals to the video frames, the system exploits the noise's high-frequency content to sharpen images and enhance dynamic range. The noise, which would normally degrade image quality, is instead used to create the appearance of higher resolution and improved visual acuity.
Solution Approach 2:
The patent changes the parameters of the video signal by superimposing noise with specific characteristics (zero-mean Gaussian white noise with controlled variance). This parameter change introduces high-frequency content that enhances perceived sharpness. The processing complexity is managed by using standard statistical properties of the noise and simple superposition operations rather than complex algorithms.
Data Source
AI summary
Displaying video in an HMD may include introducing unperceived noise to the video frame signal in order to enhance dynamic range. For example, each of a viewer's left and right eyes have a field of view (FOV) corresponding to a portion of pixels shown on the HMD. For each of these portions of pixels, the VR system may combine a noise signal (e.g., zero-mean Gaussian white noise) with the video signals corresponding to each of the portions of pixels. The introduction of such noise may improve the dynamic range of the viewer. Further, in some implementations, the noise signal that is combined with the left video signal may be slightly different from the noise signal that is combined with the right video signal. Such slightly different noise signals may provide further improvement to the image seen by the viewer due to binocular summation.