Digital Image Noise Reduction via Optimal Primary Color Signal Mixing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image noise reduction methods in digital cameras, particularly in high ISO modes, face challenges in preserving image details due to improper treatments of image characteristics, leading to limited noise reduction effectiveness and high computational requirements.
Innovation Solution
An image noise reduction method based on an optimal primary-color signal, where the digital image is decomposed into high-frequency and low-frequency signals using filters, and the optimal signal is used to amend the worse signal by mixing high-frequency components, while setting a weight ratio to maintain image details and reduce noise spots.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gain compensation is applied to each primary color to achieve optimal signal strength, then the signal-to-noise ratio of primary colors is improved, but the amplification ratios of noises become different for each primary color
Solution Approach 1:
The patent changes the parameter of gain compensation by introducing a noise reduction gain that is smaller than the original gain compensation. This modified gain is applied selectively to primary colors with larger noise, thereby reducing their noise amplitude while maintaining signal strength, thus achieving more uniform noise levels across all primary colors.
2Object-generated harmful factors
If noise reduction processing is performed in the color difference domain, then noise spots are reduced, but image details cannot be reserved due to improper treatments on image characteristics
Solution Approach 1:
The patent segments the image processing into two distinct domains: luminance signal processing and color difference signal processing. By separating these processing paths, the invention applies noise reduction selectively to the color difference domain while preserving the luminance domain details, thus reducing noise spots without losing image details.
Solution Approach 2:
The patent applies different processing qualities to different components of the image signal. The luminance signal undergoes minimal processing to preserve details, while the color difference signal undergoes noise reduction processing. This local quality approach ensures that noise spots are reduced in color information without affecting the detailed information in luminance.
3Loss of information
If image details are required to be reserved through proper treatments, then a great deal of computations is needed, but noise reduction effectiveness is limited
Solution Approach 1:
The patent extracts only the necessary noise reduction processing from the complete image processing pipeline. By taking out just the color difference domain noise reduction step and applying it selectively, the invention achieves noise reduction effectiveness without requiring extensive computational resources for full-image processing.
Data Source
AI summary
An image noise reduction method based on an optimal primary-color signal is suitable for a digital camera to reduce noises in a digital image. The image noise reduction method mainly includes the following steps, taking out primary-color signals of a digital image, including taking out high-frequency and low-frequency signals of the optimal primary-color signal and high-frequency and low-frequency signals of the worst primary-color signal by using a high-pass filter and a low-pass filter; next, mixing the high-frequency signal of the optimal primary-color signal with that of the worst primary-color signal, and superposing the mixed signal to the low-frequency signal of the worst primary-color signal, so as to get a digital image signal after noise reduction.