Image Processing Patch Pixel Count Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing patch-based noise reduction methods face challenges in accurately detecting similar patches and performing noise reduction when the number of pixels in a patch is large, leading to ineffective noise reduction processes.
Innovation Solution
An image processing apparatus and method that sets a focused patch with a specific number of pixels, detects similar patches, adjusts the number of pixels for target patches, and performs noise reduction on these target patches to generate an output image, allowing for accurate noise reduction by varying the number of pixels used for detection and reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of pixels are used in a patch for noise reduction, then the noise reduction accuracy should improve, but the detection of similar patches becomes less accurate and the processing becomes less effective
Solution Approach 1:
The patent segments the patch processing into two distinct stages with different pixel counts: a first number of pixels for detecting similar patches, and a second number of pixels for performing noise reduction. This segmentation allows each stage to use the optimal patch size for its specific function, resolving the contradiction between detection accuracy and noise reduction accuracy.
Solution Approach 2:
The patent applies different quality parameters (number of pixels) to different functional requirements: using a first number of pixels optimized for similar patch detection and a second number of pixels optimized for noise reduction processing. This local differentiation of parameters allows each process to operate at its peak effectiveness without being constrained by the other's requirements.
2Loss of information
If a large number of pixels are used in a patch, then more information is available for noise reduction, but the processing time and computational complexity increase
Solution Approach 1:
The patent divides the processing workflow into two segments: first detecting similar patches using a smaller first number of pixels (reducing computational load), then performing noise reduction on selected patches using a second number of pixels. This segmentation reduces overall processing time while preserving necessary information for effective noise reduction.
Solution Approach 2:
The patent applies partial action by using a first number of pixels for detection (less than what would be used for full noise reduction) and then applying the second number of pixels only to the detected similar patches. This avoids the excessive computational burden of processing all pixels in all patches at full resolution, reducing processing time while maintaining noise reduction effectiveness.
Data Source
AI summary
There is provided with an image processing apparatus. A focused patch setting unit sets a focused patch in an input image. The focused patch has a first number of pixels. A detecting unit detects a plurality of similar patches that are similar to the focused patch in the input image. A target patch setting unit sets, based on the plurality of similar patches, a plurality of target patches including at least one pixel included in the similar patch and having a second number of pixels that is different from the first number of pixels. A noise reduction processing unit performs, based on the plurality of target patches, a noise reduction process on the target patches. A merging unit merges the target patches whose noise has been reduced by the noise reduction processing unit to generate an output image.