VR Image Deblurring via Perceptual Afterimage Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing deblurring techniques for virtual reality devices do not effectively address motion blur caused by human perception characteristics, leading to image quality deterioration due to afterimages.
Innovation Solution
A method and device that predict and calculate the cumulative afterimage effect based on user perceptual characteristics, generating blur compensation images to remove motion blur by determining a cumulative value using a blur kernel and tracking user eye movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If existing deblurring techniques are used, then motion blur from camera movement can be removed, but motion blur caused by human perception characteristics (afterimages) cannot be effectively addressed
Solution Approach 1:
The patent applies parameter changes by modifying the deblurring process to incorporate human perception parameters. Specifically, it changes the approach from purely optical deblurring to perceptual deblurring by introducing parameters related to afterimage duration, persistence, and visual system response characteristics. This allows the system to adapt the deblurring algorithm to match human visual perception, thereby resolving motion blur that previous techniques could not address.
Solution Approach 2:
The patent implements dynamics by making the deblurring process adaptive and time-varying. Instead of using static deblurring kernels, the system dynamically adjusts the deblurring parameters based on the temporal characteristics of the image sequence and estimated human visual system response. This dynamic adaptation enables the system to handle varying motion conditions and perception characteristics, improving image quality across different scenarios.
2Manufacturing precision
If deblurring is applied to remove motion blur, then image clarity improves, but artifacts may be introduced due to incorrect assumptions about blur causes
Solution Approach 1:
The patent employs feedback mechanisms by using the estimated afterimage characteristics to guide the deblurring process. The system continuously monitors the image sequence, estimates the afterimage effects based on temporal patterns and motion information, and uses this feedback to adjust the deblurring parameters in real-time. This closed-loop approach helps prevent artifacts by ensuring that the deblurring operation matches the actual perceptual blur present in the images.
Solution Approach 2:
The patent applies preliminary action by pre-calculating or pre-estimating the afterimage characteristics before performing the final deblurring operation. The system uses motion information and temporal patterns from the image sequence to predict the afterimage effects, and then applies appropriate compensation in advance. This preliminary estimation allows the deblurring process to be more accurate and less likely to introduce artifacts, as the compensation is tailored to the specific perceptual conditions.
Data Source
AI summary
Disclosed is a method, performed by a device, of processing an image, the method including: for an original image at a particular time point among a plurality of original images having a sequential relationship in terms of time, determining a cumulative value due to an afterimage of another original image before the particular time point; based on the determined cumulative value and the plurality of original images, obtaining a plurality of blur compensation images for removing a blur caused by the afterimage; and outputting the obtained plurality of blur compensation images.