Asymmetric Stereo Imaging with Depth Map Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high cost of dual-lens cameras is a significant issue due to the need for two cameras with the same specifications, which doubles the costs, while existing solutions fail to efficiently generate stereo images with high resolution while minimizing expenses.
Innovation Solution
A method and apparatus that utilize a high-resolution camera and a low-resolution camera to generate stereo images by upsampling the low-resolution image to match the high-resolution camera's resolution, generating a depth map, re-projecting the high-resolution image to reconstruct a reference image, and compensating occlusion regions, thereby creating high-resolution stereo images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dual lenses with the same specification are arranged on a camera, then stereo images with high resolution can be captured, but the costs will be doubled
Solution Approach 1:
The patent employs asymmetric camera configuration where one camera uses a high-resolution lens and the other uses a low-resolution lens, breaking the symmetry of traditional dual-lens setups. This allows the system to achieve stereo imaging functionality while reducing costs by using different specification cameras instead of identical ones
Solution Approach 2:
The patent creates a virtual copy of the high-resolution image through re-projection algorithms. The high-resolution image from one camera is re-projected to generate a reference image for the low-resolution camera's viewpoint, effectively copying the high-resolution quality to the stereo pair without requiring a second high-resolution camera
2Ease of manufacture
If a low-resolution camera is used to reduce costs, then costs are minimized, but the quality of stereo images deteriorates
Solution Approach 1:
The patent introduces depth information as an intermediary element that mediates between the low-resolution and high-resolution images. By calculating depth maps from the stereo pair and using them for re-projection, the system transfers high-resolution details to the reference image, compensating for the low-resolution camera's limitations
Solution Approach 2:
The patent changes the resolution parameter asymmetrically between the two camera channels. One camera operates at high resolution while the other at low resolution, but through computational processing (re-projection and occlusion compensation), the final stereo output achieves high resolution in both views
3Measurement precision
If high-resolution cameras are used for both left and right views, then stereo image quality is maximized, but device complexity increases
Solution Approach 1:
The patent extracts the high-resolution capture function to only one camera, while the other camera serves a dual purpose: capturing low-resolution stereo data and providing depth information through disparity calculation. This extraction of the high-resolution function reduces system complexity while maintaining stereo image quality through computational re-projection
Data Source
AI summary
A stereo image generating method and an electronic apparatus utilizing the method are provided. The electronic apparatus includes a first camera and a second camera capable of capturing stereo images, and a resolution of the first camera is larger than that of the second camera. In the method, a first image is captured by the first camera, and a second image is captured by the second camera. The second image is upscaled to the resolution of the first camera, and a depth map is generated with use of the first image and the upscaled second image. With reference to the depth map, the first image is re-projected to reconstruct a reference image of the second image. An occlusion region in the reference image is detected and compensated by using the upscaled second image. A stereo image including the first image and the compensated reference image is generated.