Adaptive Image Recovery Filter for Over-Recovery Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recovery processing technologies often result in over-recovery of image data due to mismatches between the actual imaging state and the state envisioned by the image recovery filter, leading to image deterioration such as undershoot and overshoot, especially when the object distance varies during three-dimensional object capture.
Innovation Solution
An image processing apparatus and method that selects and generates an image recovery filter based on real-time imaging information, including lens focal length, aperture value, and imaging distance, and performs convolution processing using a two-dimensional filter to correct aberrations without Fourier transforms, while incorporating change amount limiting processing to prevent over-recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If image recovery processing is performed using a fixed image recovery filter, then image quality can be improved by correcting aberrations, but over-recovery occurs when the imaging state varies, leading to image deterioration such as undershoot and overshoot
Solution Approach 1:
The patent applies dynamics by making the image recovery filter adaptive rather than fixed. The filter dynamically adjusts its parameters based on detected imaging conditions (such as focus distance, aperture, and lens position) to match the actual blur characteristics of the captured image. This allows the system to maintain high recovery accuracy across varying imaging scenarios without causing over-recovery artifacts.
Solution Approach 2:
The patent implements parameter changes by modifying the image recovery filter's characteristics according to the detected imaging state. When imaging conditions change (e.g., different focus distances or aperture values), the system recalculates and applies appropriate filter parameters that correspond to the actual blur component, thereby preventing mismatch between the filter and the image degradation pattern.
2Reliability
If the image recovery filter is changed according to imaging state, then recovery accuracy is improved, but device complexity increases due to the need to track and respond to multiple imaging parameters
Solution Approach 1:
The patent applies self-service by enabling the image recovery filter to automatically adapt to changing imaging conditions without requiring complex external control. The system uses information already available from the imaging process (such as metadata from the lens or sensor) to self-adjust the filter parameters, eliminating the need for additional complex control mechanisms while maintaining high recovery accuracy.
3Speed
If convolution processing is used instead of Fourier transforms, then processing speed is improved, but filtering precision may be reduced
Solution Approach 1:
The patent applies mechanics substitution by replacing the complex Fourier transform-based filtering mechanism with a simpler convolution-based approach. By designing a convolution filter that directly models the blur component characteristics, the system achieves comparable or superior filtering precision while significantly reducing computational complexity and processing time.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces over-recovery and enhances image quality by ensuring the image recovery filter matches the actual imaging conditions, minimizing undershoot and overshoot, and maintaining consistent pixel signal value changes even after gamma correction.
Implementation Method 1
the original image data can similarly be obtained by performing convolution processing on the image data in the actual plane as illustrated in the following equation
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
Provided is an image processing apparatus that reduces over recovery of image data by image recovery processing. An image processing means (104) selects an image recovery filter corresponding to an imaging condition and performs image recovery processing on captured image data using the image recovery filter. Further, the image processing means (104) limits a change amount of a pixel signal value of the image data by the image recovery processing based on a change amount limit value determined according to characteristics of a gamma correction processing that is applied after the image recovery processing.