Image Processing Device Gradation Correction Using Weighted Luminance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods for gradation correction, such as Retinex processing, can introduce unwanted gradients not present in the original image, and existing techniques often rely on a single set of gradation conversion characteristics or user-determined weights, which may lead to adverse effects like halo formation and inefficient correction.
Innovation Solution
An image processing method that calculates a weighted average of luminance values using multiple gradation conversion characteristics, represented by monotonically increasing and convex upward curves, to perform pixel-by-pixel gradation correction, reducing the difference between characteristics and allowing for region-specific fine-tuning, thereby minimizing adverse influences like halo formation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Retinex processing is performed for gradation correction, then local brightness correction is achieved, but unwanted gradients appear in the corrected image
Solution Approach 1:
The patent changes the parameters of gradation conversion by using multiple conversion characteristics (first and second characteristics) with different properties (one emphasizing dark portions, the other bright portions). By dynamically selecting and combining these different parameter sets based on local luminance conditions, the method achieves accurate gradation correction while avoiding the unwanted gradients that result from using a single fixed conversion parameter
Solution Approach 2:
The patent applies different gradation conversion characteristics to different local regions of the image based on their luminance properties. Dark regions use the first conversion characteristic optimized for dark portions, while bright regions use the second conversion characteristic optimized for bright portions. This local adaptation prevents the introduction of unwanted gradients by matching the conversion approach to the local image characteristics
2Device complexity
If a single set of gradation conversion characteristics is used, then processing is simple, but adverse effects like halo formation occur
Solution Approach 1:
The patent introduces dynamic selection of gradation conversion characteristics based on local luminance conditions. Instead of using a static single conversion characteristic, the system dynamically determines which characteristic (first or second) to apply based on whether the local region is dark or bright. This dynamic adaptation eliminates halo formation while maintaining reasonable processing complexity through automated region-based selection
3Adaptability or versatility
If user-determined weights are used for combining gradation characteristics, then flexibility is improved, but operation complexity increases and efficiency decreases
Solution Approach 1:
The patent implements self-service by automatically determining the weights for combining gradation conversion characteristics based on local luminance values. The system calculates the weight ratio between the first and second conversion characteristics according to the luminance level of each local region, eliminating the need for manual user input. This automated approach maintains flexibility in adapting to different image conditions while preserving operational simplicity
Solution Approach 2:
The patent uses feedback from local luminance measurements to automatically adjust the weighting of different gradation conversion characteristics. The luminance value of each local region serves as feedback that determines the appropriate weight ratio between the first and second conversion characteristics, enabling the system to adaptively optimize correction results without requiring manual intervention
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image processing device includes: a parameter calculation unit that obtains a plurality of values based on a plurality of different conversion parameters for a pixel in image data; a weighting unit that performs predetermined weighting on the plurality of values obtained by the parameter calculation unit; and a gradation correction unit that performs gradation correction of the image data based on a result of the weighting by the weighting unit.