Foreground Object Removal from Composite Images via Parallax Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The parallax effect causes foreground objects to significantly move relative to the background when images are taken from different perspectives, making it challenging to process these images into a composite image without the foreground object.
Innovation Solution
A method and apparatus that digitally process two images from different perspectives to identify obscured portions at an imaging plane, determining if these portions overlap, and outputting an indication of whether the images can be successfully processed to produce a composite image without the foreground object, allowing the user to reframe the scene if necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If images are taken from different perspectives to enable foreground object removal, then the ability to remove the foreground object is improved, but the parallax effect causes significant movement of the foreground object relative to the background making successful removal difficult
Solution Approach 1:
The system performs preliminary analysis by identifying obscured portions in each image before attempting composite image generation. This preliminary action determines whether the images are suitable for processing, allowing users to reframe if necessary, thus preventing wasted processing on unsuitable image pairs and improving overall reliability of object removal
Solution Approach 2:
The system provides feedback to users about whether the captured images will successfully allow foreground object removal based on the overlap analysis of obscured portions. This feedback mechanism enables users to adjust their拍摄 positioning to achieve better results, effectively managing the parallax effect through iterative refinement
2Reliability
If the obscured portions from two images do not overlap at the imaging plane, then the foreground object can be successfully removed, but if they overlap, the removal cannot be achieved
Solution Approach 1:
The system performs preliminary identification and comparison of obscured portions before composite image generation. By analyzing whether the obscured portions overlap at the imaging plane in advance, the system determines suitability for processing, preventing wasted computational resources and guiding users to capture better-positioned images for successful object removal
Data Source
AI summary
A method including digitally processing a first image of a scene from a first perspective, to identify a first obscured portion of the scene at an imaging plane that is obscured by a forward object in the scene; digitally processing a second image of the scene from a second perspective, to identify a second obscured portion of the scene at the imaging plane that is obscured by the forward object in the scene; and outputting, if the first obscured portion of the scene and the second obscured portion of the scene are overlapping at the imaging plane, an indication that the first image and the second image cannot be processed successfully to produce a third composite image of the scene without the forward object.