Omnidirectional Texture Image Occlusion Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
When the viewpoint for generating and reproducing an omnidirectional image differs, the generated texture image at the reproduction viewpoint includes occlusion regions, leading to degraded picture quality due to hidden objects in the rear being obscured by objects in the front.
Innovation Solution
An image processing apparatus and method that generate a texture image at a given viewpoint using a first layer image with texture and depth information from one viewpoint and a second layer image with texture and depth information from a different viewpoint, specifically handling occlusion regions by incorporating texture and depth images from both layers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a texture image is generated at a reproduction viewpoint using an omnidirectional image captured at a different viewpoint, then the viewing range of the viewer is covered, but occlusion regions appear and picture quality degrades
Solution Approach 1:
The patent divides the texture image generation process into two layers: a first layer capturing the omnidirectional image at the generation viewpoint, and a second layer capturing images at additional viewpoints. This segmentation allows the system to handle different viewing scenarios separately, using the appropriate layer to avoid occlusion artifacts while maintaining comprehensive viewing range coverage.
Solution Approach 2:
The patent introduces depth information as an intermediary element that mediates between the texture images from different viewpoints. By using depth maps, the system can identify occlusion regions and selectively replace them with corresponding regions from alternative viewpoints, thus resolving the occlusion problem without compromising the overall viewing range coverage.
2Manufacturing precision
If multiple layer images with texture and depth information are used to handle occlusion regions, then picture quality improves, but device complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-capturing omnidirectional images at multiple viewpoints and pre-computing their corresponding depth maps before reproduction. This allows the reproduction apparatus to simply select and composite the appropriate layers during playback, significantly reducing real-time processing complexity while maintaining high picture quality through multi-layer occlusion handling.
Solution Approach 2:
The patent applies local quality by processing only the occlusion regions using the second layer images, rather than processing the entire texture image. The system identifies occlusion areas through depth comparison and selectively replaces only those specific regions, thereby improving picture quality where needed while minimizing the overall processing burden and device complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to an image processing apparatus and an image processing method that make it possible to generate a texture image of high picture quality at a predetermined viewpoint using an omnidirectional image. A drawing section generates a display image of a given viewpoint using a first layer image including a texture image and a depth image of an omnidirectional image at a first viewpoint and a second layer image including a texture image and a depth image at a second viewpoint different from the first viewpoint in an occlusion region at a viewpoint of the first layer image. The present disclosure can be applied, for example, to a home server that generates a display image of a predetermined viewpoint from an omnidirectional image or the like.