Blur Estimation Using Iterative Correction Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately estimate blur in images with small signal variations, leading to deteriorated estimation accuracy, especially in areas with limited edges and textures, such as those affected by diffraction, aberration, defocus, and hand shake.
Innovation Solution
An image processing method that performs iterative calculation processing, repeating correction and estimation steps using different calculation expressions to generate multiple correction signals, improving signal variation and estimation accuracy in blur estimation areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single blurred image is used for blur estimation, then the method is simple and fast, but the estimation accuracy deteriorates in areas with small signal variations
Solution Approach 1:
The blurred image is divided into multiple estimation areas, with each area processed independently to estimate local blur characteristics. This segmentation allows the system to handle areas with small signal variations by focusing on local regions rather than relying on global image statistics alone.
Solution Approach 2:
Multiple candidate correction signals are generated in advance through iterative calculation processing before final blur estimation. This preliminary generation of multiple candidates allows the system to prepare sufficient information for accurate estimation even in challenging areas with limited signal variations.
2Measurement precision
If iterative calculation processing with multiple correction signals is performed, then the estimation accuracy improves, but the processing time increases
Solution Approach 1:
The system performs iterative calculation processing to generate multiple candidate correction signals, which is more than the minimum required. This excessive generation of candidates ensures sufficient accuracy even in areas with small signal variations, while the iterative nature allows for progressive refinement rather than requiring all calculations to be performed simultaneously.
Solution Approach 2:
The iterative calculation processing continuously refines the correction signals through repeated correction and estimation steps. This continuous refinement allows the system to progressively improve accuracy while managing computational load through staged processing rather than attempting a single complex calculation.
3Measurement precision
If the blur is estimated using statistical information from natural images, then the method works for hand shake correction, but it fails when edges are insufficient or blur varies by area
Solution Approach 1:
The system estimates blur characteristics locally for each estimation area rather than applying a uniform global estimation method. This local quality approach allows the system to adapt to varying blur conditions in different areas, handling Shift-variant blur and areas with small signal variations by processing each region with appropriate local characteristics.
Solution Approach 2:
The system dynamically adjusts the estimation process for different areas based on their characteristics. By performing iterative calculation processing that adapts to local signal variations and edge densities, the system can handle diverse blur conditions including hand shake, diffraction, aberration, defocus, and disturbance in a unified yet flexible manner.
Data Source
AI summary
An image processing apparatus includes an acquirer which acquires a blurred image, and a generator which acquires a blur estimation area of at least a part of the blurred image to generate an estimated blur based on the blur estimation area, and the generator generates the estimated blur by performing iterative calculation processing that repeats correction processing and estimation processing, the correction processing correcting a blur included in information relating to a signal in the blur estimation area to generate information relating to a correction signal, and the estimation processing estimating a blur based on the information relating to the signal and the information relating to the correction signal, and generates, as the information relating to the correction signal, information relating to a plurality of correction signals by using a plurality of different calculation expressions in at least one correction processing during the iterative calculation processing.