Adaptive Pixel Group Noise Reduction for Digital Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction methods in digital imaging struggle to balance noise reduction effectiveness with processing load, particularly when the number of pixels used for similarity calculation is increased, leading to either inaccurate noise estimation or reduced noise reduction due to limited reference pixels.
Innovation Solution
An image processing apparatus dynamically adjusts the number of pixels used for similarity calculation based on the noise amount and imaging conditions, such as ISO sensitivity and exposure time, to optimize noise reduction while minimizing processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of pixels included in the area for determining the similarity is increased, then the similarity can be calculated more accurately, but the number of reference pixels with high similarity to the target pixel is reduced
Solution Approach 1:
The patent applies dynamics by making the pixel group size adaptive rather than fixed. The determination unit dynamically selects from multiple pixel group candidates (with different numbers of pixels) based on the calculated similarity and noise characteristics. This allows the system to optimize between similarity calculation accuracy and the number of reference pixels available for noise reduction, resolving the contradiction between measurement precision and quantity of substance.
2Productivity
If a large weight is applied to a reference pixel with noise pattern similar to the target pixel, then the weighted average calculation is simplified, but the noise reduction effect deteriorates
Solution Approach 1:
The patent implements feedback by using the calculated similarity as a basis for determining weights. The determination unit calculates similarity between target and reference pixels, then uses this similarity information to assign appropriate weights. This feedback mechanism ensures that pixels with genuinely similar patterns (not just noise patterns) receive higher weights, maintaining noise reduction effectiveness while preserving processing efficiency.
3Device complexity
If noise reduction processing is performed with limited reference pixels, then the processing load is reduced, but the noise reduction effectiveness deteriorates
Solution Approach 1:
The patent applies parameter changes by varying the pixel group size (number of pixels used for similarity calculation) based on imaging conditions and noise characteristics. The determination unit selects from multiple pixel group candidates with different sizes, adjusting this parameter to optimize the balance between processing load and noise reduction effectiveness for different scenarios.
Data Source
AI summary
An image processing apparatus to execute noise reduction processing on image data includes a setting unit, a determination unit, and an output unit. The setting unit sets a pixel group from among a plurality of pixel group candidates. The plurality of pixel group candidates includes at least a first pixel group having a plurality of pixels being a first number of pixels or a second pixel group having a plurality of pixels being a second number of pixels which is different from the first number of pixels. The determination unit determines, based on a similarity between a target pixel and a reference pixel that is obtained according to the set pixel group, a weight corresponding to the reference pixel. The output unit outputs a value, calculated based on a pixel value of the reference pixel and the weight, as a noise-reduced pixel value of the target pixel.