Image Noise Reduction via Selective Pixel Averaging at Edge Corners
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction methods in images often blur edges, especially those with low contrast, and setting a small threshold to preserve edges results in reduced noise reduction effectiveness.
Innovation Solution
An image processing apparatus that sets parameters for noise reduction based on whether a pixel is in a corner portion or not, averaging pixels for noise reduction only when not in a corner, thus preserving edge sharpness while reducing noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a threshold is set large to preserve edge sharpness, then edge sharpness is maintained, but noise reduction effectiveness is reduced
Solution Approach 1:
The patent applies different noise reduction strategies to different regions of the image based on local characteristics. Corner pixels (where edges intersect) are excluded from averaging to preserve sharpness, while non-corner pixels on edges are included. This localized differentiation resolves the contradiction by adapting the noise reduction intensity to the specific geometric context of each pixel.
Solution Approach 2:
The patent segments the image processing into distinct cases: corner pixels and non-corner pixels. By identifying corner portions where two different edges intersect and treating them differently from other edge pixels, the method enables selective application of noise reduction, thereby maintaining edge sharpness while reducing noise in appropriate regions.
2Measurement precision
If a threshold is set small to detect pixels with high accuracy, then pixel detection accuracy is improved, but the number of pixels for averaging decreases, reducing noise reduction effect
Solution Approach 1:
The patent differentiates between corner pixels and non-corner pixels, applying noise reduction only to the latter. This local quality approach ensures that pixels are processed according to their specific geometric context, maintaining detection accuracy for corner pixels while enabling effective noise reduction for non-corner pixels through averaging.
Solution Approach 2:
By segmenting the pixel population into corner and non-corner categories, the patent enables selective noise reduction. The segmentation allows the system to maintain high detection accuracy for corner pixels (which are excluded from averaging) while applying noise reduction to non-corner pixels, thus resolving the trade-off between detection accuracy and noise reduction effectiveness.
3Object-affected harmful factors
If averaging is applied to all pixels, then noise reduction effect is maximized, but edge sharpness is lost
Solution Approach 1:
The patent applies averaging selectively based on the local geometric characteristics of pixels. Non-corner pixels on edges are included in averaging to reduce noise, while corner pixels (where two different edges intersect) are excluded to preserve sharpness. This local differentiation resolves the contradiction by adapting the noise reduction application to the specific context of each pixel.
Solution Approach 2:
The patent segments the image into corner regions and non-corner regions, applying different noise reduction treatments to each. This segmentation enables the system to maximize noise reduction in non-corner areas while preserving edge sharpness in corner areas, thereby resolving the contradiction between noise reduction effectiveness and edge sharpness maintenance.
Data Source
AI summary
An image processing apparatus to execute a noise reduction process on an image includes setting and calculation units. The setting unit sets a parameter for a pixel of interest in the image. The calculation unit calculates a pixel value of the pixel of interest by using the parameter. Where the pixel of interest is not a pixel in a corner portion where two different edges intersect, the calculation unit averages a plurality of pixels, thereby calculating the pixel value of the pixel of interest subjected to the noise reduction process. Where the pixel of interest is a pixel included in the corner portion, the averaging is not executed and the pixel value of the pixel of interest is calculating by treating the pixel value of the pixel of interest as it is as the pixel value of the pixel of interest subjected to the noise reduction process.