Image Restoration Apparatus Using Overlapping Block Blur Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image restoration methods fail to accurately estimate the point spread function (PSF) based on the blur level, leading to inadequate deblurring and artifacts in images, especially in mobile phone cameras where lens fixation results in varying blur levels due to subject distance.
Innovation Solution
An image restoration method and apparatus that generates overlapping blocks, estimates blur levels using unit step responses, and applies a deconvolution filter to refine the PSF, enabling accurate deblurring by locally adjusting the blur estimation and feature mapping for each block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional image restoration methods use a fixed PSF for deblurring, then the device complexity is reduced, but the measurement precision of blur level estimation deteriorates
Solution Approach 1:
The image is divided into multiple overlapping blocks, and each block is processed independently to estimate local blur levels. This segmentation allows the system to adapt to varying blur conditions across different regions without requiring complex global analysis, thus maintaining relatively simple device complexity while improving measurement precision.
Solution Approach 2:
Instead of applying a single fixed PSF to the entire image, the method estimates and applies different PSFs for different local blocks based on their specific blur levels. This local quality approach enables precise blur correction for each region while keeping the overall system complexity manageable through modular processing.
2Device complexity
If the lens is fixed without an actuator, then the device complexity and size are reduced, but the manufacturing precision of image focus deteriorates
Solution Approach 1:
The patent replaces the mechanical autofocus system (actuator and movable lens) with a digital image processing system that estimates blur levels and applies deconvolution filters. This substitution eliminates the need for complex mechanical components while achieving effective focus correction through computational methods, thus reducing device complexity while maintaining imaging precision.
3Measurement precision
If local blur estimation is performed for each block, then the measurement precision of PSF estimation is improved, but the productivity of image processing deteriorates
Solution Approach 1:
The method performs preliminary blur level estimation for each block before applying the deconvolution filter. By pre-calculating the blur characteristics and preparing appropriate PSFs in advance, the system avoids repeated complex computations during the actual deblurring process, thus maintaining high measurement precision while improving processing productivity.
Solution Approach 2:
Adjacent overlapping blocks share common computational results and PSF estimations. By merging and reusing calculations from overlapping regions, the system reduces redundant computations across blocks, thereby improving overall processing productivity while maintaining the precision benefits of local blur estimation.
Data Source
AI summary
Provided is an image restoration method and apparatus, in which an image is restored by estimating the blur level of the image. The image restoration apparatus includes a block generation module which generates a plurality of overlapping blocks for an input image; a blur estimation module which generates a plurality of unit step responses for a current overlapping block from the overlapping blocks; a feature-map generation module which generates a feature map for the current overlapping block based on the unit step responses of the current overlapping block; and a filter application module which deblurs the current overlapping block by using a deconvolution filter to filter the feature map of the current overlapping block.