Under-Display Array Camera Blending to Reduce Diffraction Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Under-display camera (UDC) technology in mobile devices suffers from limited light transmission through display panels, leading to image artifacts such as noise and light diffraction, resulting in poor image quality.
Innovation Solution
The use of multiple under-display cameras and processors to capture and align image frames, generate optical flow and occlusion maps, warp images based on these maps, and blend them to improve image quality, specifically for 3D scenes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple under-display cameras are used to capture images, then image quality is improved by reducing diffraction artifacts, but device complexity increases
Solution Approach 1:
The patent combines images from multiple under-display cameras through alignment and blending operations. The processor aligns the first and second images captured by different cameras, then blends them to produce a final image with reduced diffraction artifacts and improved quality, effectively merging multiple light paths to overcome the limitations of individual cameras under the display
Solution Approach 2:
The patent segments the imaging function across multiple cameras positioned at different locations under the display. Each camera captures a separate image that is then processed independently through alignment and blending, allowing the system to divide the imaging task among multiple components to achieve superior overall image quality
2Measurement precision
If images from multiple cameras are aligned and blended, then pixel-level correspondence is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary alignment operations on the images from multiple cameras before blending. By pre-aligning the images using feature matching and transformation operations, the system establishes accurate pixel-level correspondence in advance, which facilitates more efficient subsequent processing and reduces overall computation time
Data Source
AI summary
A first image frame and a second image frame are captured using first and second under-display cameras positioned under an LED display. The second image frame is globally aligned to the first image frame to generate a globally-aligned second image frame. An optical flow map based on the first image frame and the globally-aligned second image frame is generated, and an occlusion map based on the first image frame and the globally-aligned second image frame is generated. The globally-aligned second image frame is warped based on the optical flow map and the occlusion map to generate a warped image frame. The first image frame and the warped image frame are blended to generate an output image.