Virtual Image Ghosting Removal via Pixel Transparency Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for creating virtual images from reference view images introduce areas of false transparency, known as ghost regions, which reduce the quality of virtual images and hinder their use in VR/AR applications by allowing background to shine through the object.
Innovation Solution
A method for setting transparency and color values in virtual images by determining transparency information components from reference view images, classifying pixels based on foreground and background information, and selecting binary transparency values to avoid ghosting effects, while ensuring accurate blending with new backgrounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If linear combination techniques are used to calculate pixel colors from reference view images, then virtual images can be created from multiple viewpoints, but ghost regions appear at object edges causing false transparency
Solution Approach 1:
The patent segments the transparency calculation process by separately determining foreground and background transparency components for each pixel. Instead of using a single linear combination formula, the method divides the transparency information into distinct foreground (object) and background components, allowing independent processing and classification to eliminate ghost regions while maintaining viewpoint flexibility.
2Productivity
If transparency values are calculated using linear combination of reference images, then virtual images can be synthesized, but ghost regions reduce image quality and VR/AR immersion
Solution Approach 1:
The patent applies local quality by classifying each pixel individually based on its foreground and background transparency components. Different transparency values are assigned to different pixel regions: pixels with dominant foreground components receive different transparency treatment than pixels with dominant background components. This local differentiation eliminates ghost regions at object edges while maintaining overall virtual image generation capability.
3Adaptability or versatility
If chroma key backgrounds are used in reference images, then virtual images can be used on different graphics backgrounds, but ghost regions cause background to shine through the object
Solution Approach 1:
The patent extracts and separates the foreground and background transparency components for each pixel from the chroma key reference images. By taking out the background transparency information and processing it separately from the foreground object transparency, the method enables clean compositing on different graphics backgrounds without the background shining through the object via ghost regions.
Data Source
AI summary
A method is provided for selecting a transparency setting and color values of pixels in a virtual image. The virtual image can be formed by combining reference images taken at different angles to produce the virtual image that views an object at a new, uncaptured angle. The method includes determining for each pixel of the virtual image, what information it carries from the reference view images. The information of the pixel is used to define a pixel category, and the category is used to select, based on logical conditions, what information will be displayed by the pixel and to set the color of the pixels.