Image Processing Noise Reduction via Vector Signal Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems face challenges in reducing noise effectively, particularly with spike noise, as current methods either require lengthy processing times or result in hue changes, and setting evaluation functions for noise reduction is unstable and time-consuming.
Innovation Solution
An image processing system that extracts local regions from an image signal, vectorizes color signals, sets weighting factors based on the target region's vector signal, and performs filtering to reduce noise, allowing for accurate and efficient noise reduction without hue changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If median filtering process is performed while enlarging the block size sequentially to reduce spike noise, then spike noise reduction is improved, but processing time becomes long and unstable
Solution Approach 1:
The patent changes the parameter of block size from dynamic sequential enlargement to fixed predetermined size. This allows the filtering process to be performed efficiently with consistent processing time while still achieving effective spike noise reduction through the fixed-size block median filtering operation.
2Object-affected harmful factors
If median filtering process is performed independently for each color signal to reduce spike noise, then spike noise reduction is improved, but hue changes occur sharply
Solution Approach 1:
The patent merges the filtering operations for multiple color signals by performing median filtering on corresponding pixels across different color channels simultaneously within the same block. This combined approach preserves hue stability while effectively reducing spike noise, as the filtering result for each color signal is derived from the same spatial neighborhood operation.
3Measurement precision
If evaluation function is set automatically with recurrent processing to reduce general noise, then noise reduction accuracy is improved, but processing time becomes unstable and elongated
Solution Approach 1:
The patent performs preliminary classification to determine the presence of spike noise before executing the filtering process. This preliminary action allows the system to select the appropriate filtering strategy in advance, avoiding unnecessary recurrent processing and evaluation function optimization, thereby achieving both accurate noise reduction and stable processing time.
4Object-affected harmful factors
If noise reducing process is performed for each color signal to reduce random noise, then random noise reduction is improved, but processing time is elongated
Solution Approach 1:
The patent merges the noise reduction operations for multiple color signals by performing the filtering process simultaneously on corresponding pixels across different color channels within the same block. This combined approach reduces processing time while maintaining effective random noise reduction for each color signal.
Data Source
AI summary
An image processing system includes an extraction unit for extracting a local region which contains one target region and at least one or more neighborhood regions from an image signal formed of a plurality of color signals from a CCD, a vectorizing unit for obtaining vector signals which contain the plurality of the color signals as components for each of the target and the neighborhood regions, a weighting set unit for setting weighting factors with respect to each of the vector signals in the local region based on the vector signal in the target region, a filtering unit for performing a filtering process using the weighting factors with respect to the vector signals in the local region, and a noise reducing unit for reducing the noise of the vector signal in the target region based on the vector signal through the filtering process.