Processing-Target Image Reprojection for Flexible Virtual Viewpoints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image generation devices lack flexibility in adjusting the viewpoint conversion image, particularly when using cylindrical space models, as they cannot display images projected on inner surfaces and require re-generation of the image when adjusting the viewpoint or scale, limiting the positioning and angle of virtual cameras.
Innovation Solution
A processing-target image generation device and method that projects input images onto a space model with adjustable coordinates, allowing for flexible adjustment of the output image by re-projecting images onto different planes and using quaternions for coordinate conversion, enabling intuitive perception of positional relationships and distances without gimbal lock issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If an image is projected onto a cylindrical space model with a fixed virtual camera position, then the viewpoint conversion image can be generated, but the virtual camera position is limited and cannot display images projected on inner surfaces
Solution Approach 1:
The patent divides the cylindrical space model into multiple coordinate systems (first coordinate system for the cylinder, second coordinate system for the processing target image). This segmentation allows independent transformation between coordinate systems, enabling flexible virtual camera positioning without regenerating the entire image structure.
Solution Approach 2:
The patent introduces adjustable scaling factors (first scaling factor for the curved surface area, second scaling factor for the plane surface area) that allow dynamic adjustment of the processed image. This enables the virtual camera position and viewing parameters to be changed flexibly without regenerating the base projection, resolving the contradiction between adaptability and complexity.
2Adaptability or versatility
If the viewpoint or scale of the output image is adjusted, then the output image can be optimized, but the processing target image must be regenerated, losing already processed information
Solution Approach 1:
The patent performs the complex projection onto the cylindrical space model and coordinate transformations in advance, storing the processed coordinates and scaling factors. When output image adjustments are needed, only the final rendering step needs to be updated, not the entire processing pipeline, thus avoiding time loss from regeneration.
Solution Approach 2:
The patent uses dynamic scaling factors that can be adjusted independently for different areas (curved surface vs. plane surface) of the processed image. This allows real-time optimization of output images for different viewing requirements without regenerating the base processing target image, saving time while maintaining flexibility.
3Measurement precision
If a cylindrical space model is used for viewpoint conversion, then the three-dimensional space representation is achieved, but the positioning and angle of virtual cameras are restricted
Solution Approach 1:
The patent introduces a processing target image plane as an intermediary between the cylindrical space model and the final output image. This intermediate plane allows coordinate transformations and scaling adjustments that decouple the fixed cylindrical geometry from the flexible virtual camera positioning, maintaining measurement precision while increasing positioning freedom.
Solution Approach 2:
The patent changes the parameter representation by introducing scaling factors and coordinate system transformations. Instead of being constrained by the fixed cylindrical coordinates, the system uses adjustable parameters (scaling factors, coordinate transformations) to achieve accurate positional relationships with flexible camera positioning.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A processing-target image generation device (100) generates a processing-target image which is an object to be subjected to an image conversion process for acquiring an output image based on an input image taken by an image-taking part. A coordinates correspondence part (10) causes input coordinates, spatial coordinates, and projection coordinates to correspond to each other, the input coordinates being on an input image plane on which the input image is located, the spatial coordinates being on a space model (R1, R2) on which the input image is projected, the projection coordinates being on a processing-target image plane (R3) on which the processing-target image is positioned and the image projected on the space model is re-projected.