Image Correction Apparatus Edge Contrast Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image correction techniques using the Retinex theory struggle to achieve sufficient contrast enhancement for objects with varying illuminance, as they preserve both illuminance edges and fine object edges, leading to insufficient contrast improvement and potential artifacts.
Innovation Solution
An image correction apparatus generates a reduced image and performs positive-side and negative-side limiting smoothing processes to differentiate between illuminance edges and object edges, calculating weighted reference values for interpolation to create a smoothed image that preserves edges while enhancing contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a smoothing process is performed using a reduced image to reduce calculation amount, then the calculation complexity is reduced, but fine object edges are also smoothed along with illuminance edges, leading to insufficient contrast enhancement
Solution Approach 1:
The patent segments the smoothing process into two distinct operations: first performing smoothing on a reduced image to obtain illuminance distribution, then applying this illuminance map to the full-resolution original image. This segmentation allows the calculation to be performed at lower resolution while preserving fine details in the final output, resolving the contradiction between calculation complexity and contrast enhancement precision.
Solution Approach 2:
The patent introduces a dimensional separation by working in two different resolution spaces: a reduced dimension for calculating illuminance distribution and the original full dimension for applying corrections. This dimensional approach allows efficient calculation in the reduced space while maintaining high-fidelity results in the original space, effectively resolving the contradiction.
2Measurement precision
If conventional smoothing is applied to preserve illuminance edges, then illuminance distribution accuracy is improved, but fine object edges are also preserved, reducing the difference between original and smoothed images and insufficiently enhancing contrast
Solution Approach 1:
The patent separates the illuminance estimation function from the detail preservation function by performing smoothing only on the reduced image to capture illuminance edges, then applying this illuminance map to the original image. This segmentation ensures that illuminance edges are accurately represented while fine object edges in the original image are preserved, maximizing the difference between original and corrected images for effective contrast enhancement.
Solution Approach 2:
The patent applies different processing qualities to different aspects of the image: low-resolution smoothing for illuminance estimation and high-resolution processing for detail preservation. This local quality approach allows accurate illuminance distribution capture without compromising fine object edge preservation, resolving the contradiction between illuminance accuracy and contrast enhancement.
3Manufacturing precision
If the number of pixels is increased for high-definition imaging, then image quality is improved, but the calculation amount for image correction increases significantly
Solution Approach 1:
The patent resolves the productivity issue by performing calculations in a reduced dimensional space (lower resolution) for illuminance estimation, then mapping results back to the high-definition original image. This approach maintains high image quality output while significantly reducing the calculation burden, effectively decoupling image quality from computational complexity.
Solution Approach 2:
The patent segments the processing into two stages: calculation-intensive operations performed on a reduced image and detail-preserving operations applied to the full-resolution image. This segmentation allows high-definition image correction while minimizing calculation amount by performing the most computationally intensive smoothing operations at lower resolution.
Data Source
AI summary
An image correction apparatus generates, for each pixel of a reduced image generated from an input image, a first smoothed image using each reference pixel in a filter area in which a difference obtained by subtracting a luminance value of the pixel from a luminance value of the reference pixel becomes less than a predetermined value, generates a second smoothed image using each reference pixel in the filter area in which the difference obtained by subtracting the luminance value of the reference pixel from the luminance value of the pixel becomes less than the predetermined value, and generates a smoothed image for correction of the input image based on the first smoothed image and the second smoothed image.