Color Similarity Evaluation Using RGB Filtering and CIEDE2000
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high computational cost of the CIEDE2000 metric for color similarity evaluation limits its application in scenarios with real-time constraints and limited hardware resources, particularly in image processing and computer vision, where color-based algorithms are affected by illumination and struggle with discrimination between similar colors.
Innovation Solution
A method and device that employ filtering modules to reduce the number of full CIEDE2000 metric calculations by using simpler properties such as the difference vector, angle, and lightness in the RGB color space, allowing for faster color similarity decisions while maintaining accuracy, and derive thresholds for these properties using regression models to filter out color pairs with low perceptual differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the CIEDE2000 metric is used for color similarity evaluation, then color discrimination accuracy is improved, but computational cost increases
Solution Approach 1:
The patent segments the color difference evaluation process into multiple stages: first computing simple Euclidean distances in RGB space to filter out obviously different color pairs, then applying the full CIEDE2000 metric only to pairs that pass the initial filter. This hierarchical segmentation reduces the number of expensive CIEDE2000 calculations while maintaining accurate color discrimination for similar colors.
Solution Approach 2:
The patent performs preliminary filtering using computationally inexpensive properties (RGB Euclidean distance, lightness difference) before applying the full CIEDE2000 metric. By pre-computing these simple metrics and using them to eliminate obviously different color pairs, the system prepares the data in advance to avoid unnecessary complex calculations, thereby reducing overall computational cost while preserving accuracy for borderline cases.
2Measurement precision
If the CIEDE2000 metric is used for color similarity evaluation, then perceptual color difference measurement is improved, but processing time increases
Solution Approach 1:
The evaluation process is divided into time-efficient filtering stages and accuracy-critical measurement stages. The initial filtering using RGB distance and lightness differences quickly eliminates dissimilar color pairs, reducing the time spent on full CIEDE2000 computations. Only color pairs that require precise perceptual measurement undergo the complete CIEDE2000 calculation, thereby optimizing processing time while maintaining measurement precision for relevant cases.
Solution Approach 2:
The patent applies partial action by using simplified metrics (RGB Euclidean distance, lightness difference) for the majority of color pairs that are obviously different, and reserves the full CIEDE2000 metric only for pairs where precise perceptual measurement is necessary. This selective application of computational effort reduces overall processing time while ensuring that perceptual color difference measurement accuracy is maintained for color pairs that matter most.
3Productivity
If simpler color difference metrics are used, then processing speed is improved, but color discrimination accuracy deteriorates
Solution Approach 1:
The patent creates a segmented evaluation pipeline where simple metrics (RGB Euclidean distance, lightness difference) handle the bulk of color pairs quickly, while the full CIEDE2000 metric is applied selectively to pairs that require accurate discrimination. This segmentation allows the system to achieve high processing speed for obvious cases while maintaining color discrimination accuracy for similar colors that require nuanced evaluation.
Solution Approach 2:
The patent introduces intermediary filtering metrics (RGB distance, lightness difference) that act as mediators between the simple fast metrics and the accurate but slow CIEDE2000 metric. These intermediaries efficiently filter out obviously different color pairs, allowing the system to maintain high processing speed while ensuring that color discrimination accuracy is preserved for pairs that pass through the intermediary filters and require full CIEDE2000 evaluation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method and device configured for color similarity evaluation between a first and a second color, comprising the steps of: calculating one or more filters from said first and second colors; determining that the first and second colors are dissimilar if at least one of the calculated filters is above a threshold predetermined for each filter; otherwise, calculating the CIEDE2000 color-difference and determining that the first and second colors are dissimilar if the calculated CIEDE2000 color-difference is above a predetermined threshold; otherwise, determining that the first and second colors are similar. A module filter may calculate the module of the difference vector between the first and second colors represented as 3D coordinates in the RGB color space, the angle between the two vectors of the first and second colors represented as 3D coordinates in the RGB color space, or the lightness difference between the first and second colors in the CIELAB color space.