Neural Scene Rendering Occlusion Correction via Adjacent View Warping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image rendering technologies face challenges in effectively handling occlusion areas during image warping, leading to poor image quality and significant warping errors.
Innovation Solution
A method that utilizes a neural scene representation (NSR) model to obtain a target image and an adjacent image, then corrects the target image by detecting occlusion areas using a visibility map and replacing occluded pixels with corresponding pixels from the adjacent image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If image warping is used for rendering, then rendering speed is improved, but warping errors increase and image quality deteriorates
Solution Approach 1:
The patent segments the image into multiple depth layers (foreground, midground, background) and applies different warping strategies to each layer. This allows critical regions to be handled with higher precision while maintaining overall rendering speed, directly resolving the contradiction between speed and precision.
Solution Approach 2:
The patent performs preliminary depth estimation and occlusion detection before the warping operation. By pre-identifying occluded regions and depth-prone areas, the system can apply corrective measures during warping, preventing errors rather than correcting them afterward, thus maintaining both speed and precision.
2Manufacturing precision
If traditional rendering methods are used, then image accuracy is maintained, but rendering efficiency decreases
Solution Approach 1:
The patent applies traditional accurate rendering methods only to critical regions identified by depth analysis and occlusion detection, while using faster warping techniques for non-critical areas. This partial application of high-precision methods maintains image accuracy where needed while preserving overall rendering efficiency.
Solution Approach 2:
The patent implements local quality enhancement by applying different rendering quality levels to different regions of the image. High-accuracy traditional rendering is applied to regions with high depth variation or occlusion risk, while lower-accuracy fast warping is used in stable regions, optimizing the balance between accuracy and efficiency.
3Productivity
If occlusion areas are not handled properly during warping, then rendering speed is maintained, but image quality and reliability deteriorate
Solution Approach 1:
The patent performs preliminary occlusion detection and depth map generation before warping operations. By pre-identifying occluded regions, the system can apply specific handling strategies during warping, ensuring reliable image quality in occlusion-prone areas without sacrificing rendering speed.
Solution Approach 2:
The patent introduces depth maps and occlusion masks as intermediary data structures that guide the warping process. These intermediaries enable the system to identify and specially handle occlusion areas during warping, maintaining both speed and reliability by directing computational resources only where needed.
Data Source
AI summary
A rendering method and rendering device are provided. The rendering method includes obtaining a target image corresponding to a target view by inputting parameter information corresponding to the target view to a neural scene representation (NSR) model, determining an adjacent view that satisfies a predetermined condition with respect to the target view, obtaining an adjacent image corresponding to the adjacent view by inputting parameter information corresponding to the adjacent view to the NSR model, and obtaining a final image by correcting the target image based on the adjacent image.