Multi-View Image Noise Reduction via Region-Specific Epi-Polar Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques face challenges in reducing noise in multi-view images while generating three-dimensional images, particularly in dynamic scenes where motion blur and depth of field compromise image quality, leading to weak light and increased noise.
Innovation Solution
An adaptive weight-based method is employed to reduce noise in multi-view images by dividing the image into flat and non-flat regions, using average pixel values in flat regions and center-view image values in non-flat regions, with adaptive weights generated based on scaled standard deviations to optimize noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If a small aperture and long exposure are used to increase depth of field, then depth of field is improved, but motion blur increases and image quality deteriorates
Solution Approach 1:
The patent divides the image processing into multiple stages: capturing multiple pinhole images with different apertures, performing initial noise reduction on each image, then combining them through multi-view synthesis. This segmentation allows each stage to optimize for its specific function without compromising overall image quality.
Solution Approach 2:
The patent changes optical parameters by using multiple pinhole cameras with different aperture sizes (small aperture for deep depth of field, large aperture for bright images). By varying the aperture parameter across multiple cameras and combining their outputs, the system achieves both deep depth of field and high image quality simultaneously.
2Manufacturing precision
If multiple pinhole cameras with small aperture and short exposure are used to minimize optical defocus and motion blur, then image quality is improved, but incoming light becomes weak and noise increases
Solution Approach 1:
The patent merges multiple pinhole images captured by different cameras into a synthesized multi-view image. By combining the signals from multiple cameras, the system achieves signal averaging that reduces noise while maintaining the benefits of small aperture settings.
Solution Approach 2:
The patent creates multiple copies of the scene from different camera viewpoints and combines them. Each camera captures a copy of the scene with optimal settings, and the synthesis process integrates these copies to produce a final image with reduced noise and enhanced quality.
3Object-affected harmful factors
If multi-view images are used for noise reduction, then noise is reduced, but computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary noise reduction on individual pinhole images before combining them into multi-view images. This preliminary processing simplifies the subsequent synthesis operation by reducing the noise content that would otherwise require more complex processing to eliminate.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image based on local characteristics. In regions with high reliability, simple averaging is used, while in regions with ambiguity or occlusion, more sophisticated multi-view synthesis is applied, optimizing computational resources.
4Object-affected harmful factors
If multi-view images are used for noise reduction, then noise is reduced, but processing speed decreases
Solution Approach 1:
The patent performs preliminary noise reduction and preprocessing on individual images before the multi-view synthesis stage. This preliminary action reduces the computational burden during the final synthesis operation, thereby improving processing speed without sacrificing noise reduction effectiveness.
Data Source
AI summary
An image processing apparatus is coupled to a plurality of image capturing devices. The image processing apparatus reduces a noise in an epi-polar image while generating a three-dimensional image from a multi view image. The image processing apparatus divides the multi view image into a flat region and a non-flat region, generates the epi-polar image from the multi view image, replaces an epi-polar line in the epi-polar image corresponding to the flat region with an average pixel value of the multi-view image, and replaces an epi-polar line in the epi-polar image corresponding to the non-flat region with a pixel value of a center-view image obtained from a centrally located image capturing device among the plurality of image capturing devices.