Sparse Dual-Pixel Interpolation for Disparity Map Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dual-pixel image sensors with sparse dual-pixels, like super photodiode sensors, face challenges in generating high-quality disparity maps for autofocus due to reduced dual-pixel density, leading to degraded image quality and disparity estimation accuracy.
Innovation Solution
An image processing apparatus and method that generates interpolated images from dual-pixel data using a processor, averaging pixel values from pairs of dual-pixels and performing bit-shift operations to create high-resolution disparity maps, maintaining the aspect ratio and pixel distribution similar to the original image sensor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of dedicated dual-pixel is increased to improve disparity detection accuracy, then autofocus performance is improved, but image quality deteriorates due to reduced typical pixels
Solution Approach 1:
The patent applies partial action by using only some pixels as dual-pixels (e.g., 1 in 16 or 1 in 32) rather than making all pixels dual-pixels. This sparse dual-pixel configuration provides sufficient disparity information for autofocus while preserving the majority of pixels for high-quality image capture, thus resolving the contradiction between disparity detection accuracy and image quality
2Ease of operation
If conventional reordering method is used to generate dual-image from sparse dual-pixels, then processing is simple, but disparity resolution and estimation accuracy are degraded
Solution Approach 1:
The patent introduces an intermediary process of generating a virtual dual-image through interpolation between adjacent dual-pixel locations. This virtual dual-image serves as a mediator that preserves high disparity resolution and estimation accuracy while working with the sparse dual-pixel configuration, avoiding the need for complex pixel reordering operations
Data Source
AI summary
An apparatus and method with image processing for sparse dual-pixel image data are included. In one general aspect, an apparatus includes a processor configured to generate a first interpolated image by performing interpolation on a first image and generate a second interpolated image by performing interpolation on a second image, wherein the first image and the second image comprise pixel values obtained by dual-pixels of an image sensor, and generate a disparity map based on the first interpolated image and the second interpolated image.