Virtual Viewpoint Image Generation with Selective Object Transparency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating virtual viewpoint images from multiple camera perspectives often result in insufficient representation capability and image quality, particularly when objects of interest are obstructed by other objects, limiting the user's ability to set the virtual viewpoint at desired angles.
Innovation Solution
An image processing apparatus that obtains images from multiple cameras, determines viewpoint information, and generates virtual viewpoint images with enhanced transparency of objects that obstruct the object of interest, allowing the object of interest to be viewed clearly even when obstructed by other objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the virtual viewpoint is set at a specific position to improve representation capability, then the object of interest can be viewed clearly, but the image quality at other angles deteriorates and the setting range is limited
Solution Approach 1:
The patent applies local quality by making only the obstructing objects transparent while keeping other objects opaque. This selective transparency allows the object of interest to be viewed clearly from any angle without sacrificing the overall image quality and realism of the scene. The transparency parameter is applied locally to specific objects based on their spatial relationship to the object of interest.
Solution Approach 2:
The patent implements dynamics by allowing the transparency of objects to change dynamically based on whether they obstruct the object of interest. As the virtual viewpoint moves or as objects move, the transparency status of objects is adjusted in real-time to maintain optimal visibility of the object of interest while preserving image quality elsewhere.
2Adaptability or versatility
If multi-viewpoint images are used to generate virtual viewpoint images, then the viewing angle flexibility is improved, but objects can still obstruct the object of interest making the viewpoint setting ineffective
Solution Approach 1:
The patent uses transparency changes (an optical property similar to color changes) to make obstructing objects transparent. This allows the virtual viewpoint images to maintain their flexibility in viewing angles while ensuring that the object of interest remains reliably visible regardless of the viewpoint position or object positions.
Solution Approach 2:
The patent introduces transparency as an intermediary property that mediates between the obstructing objects and the object of interest. By adjusting the transparency of intermediate objects, the system maintains the effectiveness of viewpoint settings while preserving the natural appearance of the scene for non-obstructing objects.
3Manufacturing precision
If the setting range of virtual viewpoint is limited to avoid poor quality images, then image quality is maintained, but the user cannot set the virtual viewpoint at desired angles
Solution Approach 1:
The patent changes the transparency parameter of objects dynamically based on their obstruction relationship with the object of interest. This parameter change allows users to set virtual viewpoints at any desired angle without compromising image quality, as the transparency adjustment compensates for potential obstructions and maintains visual clarity throughout the entire viewpoint range.
Data Source
AI summary
An image processing apparatus includes: an image obtaining unit configured to obtain images based on image capturing by a plurality of cameras capturing an imaging area in a plurality of directions; an information obtaining unit configured to obtain viewpoint information that indicates a virtual viewpoint; and a generation unit configured to generate a virtual viewpoint image based on the images obtained by the image obtaining unit and the viewpoint information obtained by the information obtaining unit, so that in the virtual viewpoint image, transparency of an object at a position that shields an object of interest determined among a plurality of objects positioned in the imaging area is higher than transparency of an object at a position that does not shield the object of interest.