Image Deblurring via Spatial Masking and Patch Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image deblurring techniques face challenges in accurately recovering latent sharp images and point spread functions due to issues like noise, outliers, and computational complexity, especially when dealing with high-resolution images from modern digital cameras.
Innovation Solution
A method involving the computation of spatial masks to suppress ringing artifacts, combined with iterative processes using modified blurry images and regularized least squares cost functions, allows for efficient deblurring of high-resolution images by downscaling and upsampling image components in the YCbCr color space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sophisticated deconvolution methods are used for high quality deblurring, then image quality is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The image is divided into multiple patches, and the deconvolution process is performed independently on each patch. This segmentation allows the complex global deconvolution problem to be broken down into simpler local problems, reducing overall computational complexity while maintaining image quality.
Solution Approach 2:
The patent extracts and removes outliers from the image data before performing deconvolution. By identifying and excluding corrupted pixel values that violate the convolution blurring model, the algorithm avoids being misled by these outliers, thereby improving accuracy without requiring excessively complex robustness mechanisms.
2Measurement precision
If clipping is applied to pixels containing extremely bright areas, noise, and outliers to reduce their influence, then accuracy of latent sharp image is improved, but severe ringing artifacts are produced
Solution Approach 1:
The patent converts the harmful effect of outliers into a beneficial detection mechanism. By measuring how much each pixel violates the convolution blurring model, the algorithm identifies outliers and applies targeted masking. This transforms the problem of dealing with corrupt data into an advantage for improving image reconstruction accuracy while avoiding ringing artifacts through selective rather than aggressive clipping.
3Measurement precision
If deconvolution is performed on high resolution images, then image quality is improved, but processing time becomes too slow for commercial use
Solution Approach 1:
The high-resolution image is divided into smaller patches for parallel processing. This segmentation enables the deconvolution algorithm to work on manageable subsets of the image simultaneously, significantly reducing processing time while maintaining the quality benefits of high-resolution deconvolution.
Solution Approach 2:
The algorithm applies deconvolution selectively to patches that contain sufficient reliable edge and texture information, rather than uniformly processing the entire image. This partial action approach focuses computational resources on regions where deconvolution will be most effective, improving processing efficiency for high-resolution images.
Data Source
AI summary
Deblurring a blurry image (14) includes the steps of (i) computing a spatial mask (256); (ii) computing a modified blurry image (264) using the blurry image (14) and the spatial mask (256); and (iii) computing a latent sharp image (16) using the modified blurry image (264) and a point spread function (260). Additionally, the image (714) can be analyzed to identify areas of the image (714) that are suitable for point spread function estimation. Moreover, a region point spread function (1630) can be analyzed to classify the point spread function(s) as representing (i) motion blur, (ii) defocus blur, or (iii) mixed motion blur and defocus blur. A point spread function (2670) can also be estimated.