Image Correction Device Adaptive Region Division Blur
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for correcting motion blur in images, such as those using high-speed shutters or multiple consecutively captured images, are ineffective in low illumination environments due to reduced accuracy and noise amplification, leading to inaccurate blur correction.
Innovation Solution
An image correction apparatus that divides input images into adaptive regions based on common blur characteristics, calculates point spread functions for each region, and interpolates these functions to correct the image, thereby reducing blur and noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-speed shutter is used to capture images for motion blur correction, then motion blur can be corrected using pixel correlation among multiple images, but in low illumination environments the captured images become dark resulting in reduced pixel correlation accuracy
Solution Approach 1:
The image is divided into multiple regions based on motion blur characteristics. Each region is processed independently to calculate region-specific PSFs, allowing adaptive correction that accounts for spatial variations in blur while maintaining accuracy in low illumination environments.
Solution Approach 2:
Different regions of the image are assigned different PSFs calculated from local pixel correlations. This local quality approach ensures that each region receives correction tailored to its specific blur characteristics, improving overall correction accuracy without requiring high image brightness.
2Illumination intensity
If sensitivity is increased to compensate for insufficient light in low illumination environments, then more light can be captured, but noise signals such as dark-current noise and thermal noise are amplified reducing pixel correlation accuracy
Solution Approach 1:
By segmenting the image into regions and calculating PSFs locally, the method reduces the impact of noise on pixel correlation. The regional approach allows for more robust statistical estimation of blur characteristics even when individual pixels are noisy due to high sensitivity settings.
3Device complexity
If a single PSF is calculated for the entire image to correct motion blur, then the correction process is simplified, but motion blur cannot be accurately corrected when blur characteristics vary across different image regions
Solution Approach 1:
The image is divided into multiple regions, and a separate PSF is calculated for each region based on local pixel correlations. This segmentation enables accurate correction of spatially varying motion blur while maintaining a relatively simple overall process structure.
Solution Approach 2:
The PSF is made dynamic and adaptive to different regions of the image rather than being static and uniform. Each region's PSF is calculated dynamically based on local image characteristics, allowing the correction system to adapt to varying blur conditions across the image.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An input image containing multiple kinds of blur is corrected without using multiple images, resulting in a target image with less blur than the input image. An image correction apparatus (100) for correcting an input image to generate a target image with less blur than the input image is provided, which includes: an adaptive region division unit (110) that divides the input image into a plurality of adaptive regions based on pixel values of pixels included in the input image; a PSF interpolation unit (120) that interpolates a point spread function for a pixel located between representative pixels each representing a corresponding one of the adaptive regions, using a PSF representing characteristics of image blur calculated for each of the adaptive regions; and an image correction unit (130) that corrects the input image using the point spread functions available after the interpolation, so as to generate the target image.