Super Resolution Image Synthesis via Virtual Camera Position Shifts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for changing the focus and depth of field in captured images using multiple cameras result in low-quality high-resolution images due to insufficient position shifts and ringing artifacts, limiting the effectiveness of super resolution processing.
Innovation Solution
An image processing method that inputs multiple images from different viewpoints, sets and selects synthesis parameters to generate higher-resolution images, and applies super resolution processing while controlling depth of field, using a system with multiple capturing devices and a digital signal processor to align and combine images effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If super resolution processing is executed on multiple camera images, then image resolution is improved, but image quality deteriorates due to ringing artifacts and insufficient position shifts
Solution Approach 1:
The patent changes the parameter of position shift by introducing a virtual camera that generates synthetic images with controlled position shifts. The virtual camera parameters (position, focal length, aperture) are adjusted to create optimal position shifts between input images, enabling effective super resolution processing without ringing artifacts. This resolves the contradiction by transforming the insufficient position shift condition into an optimal condition through parameter modification.
2Volume of moving object
If multiple compact cameras are used for depth of field control, then device size is reduced, but image resolution deteriorates due to small pixel count
Solution Approach 1:
The patent merges multiple low-resolution images from compact cameras with a virtual camera to synthesize a high-resolution image. By combining the actual captured images with synthetically generated images that have different position shifts and perspectives, the system achieves high resolution output while maintaining compact device size. This merging approach allows the system to overcome the pixel count limitation of individual compact cameras.
Solution Approach 2:
The virtual camera acts as an intermediary that generates synthetic images to supplement the actual captured images. These intermediary images provide the necessary position shifts and perspectives that compact cameras cannot capture individually, enabling super resolution processing while maintaining the compact form factor. The virtual camera mediates between the limited capabilities of compact cameras and the requirement for high-resolution output.
3Reliability
If images are shifted and superposed for depth of field control, then depth of field is improved, but position shift information is lost
Solution Approach 1:
Instead of shifting images to achieve depth of field control (which loses position shift information), the patent inverts the approach by using the known camera parameters to calculate and generate images with specific position shifts. The virtual camera computes what the images would look like from different positions and perspectives, then uses these generated images for both depth of field control and super resolution processing. This inversion preserves and utilizes position shift information rather than losing it.
Data Source
AI summary
A plurality of first image data having a first resolution, which are obtained by capturing images from a plurality of viewpoints, and capturing information in the capturing operation are input. Based on the capturing information, a plurality of candidate values are set as a synthesis parameter required to synthesize second image data having a second resolution higher than the first resolution from the first image data. Using a candidate value selected from the plurality of candidate values as a synthesis parameter, the second image data is synthesized from the plurality of first image data.