Determining the colorimetric attributes of multicolored materials based on a global correction and unsupervised image...

Area of Science:

• Color Science
• Computer Vision
• Image Processing

Background:

• Accurate color measurement of multicolored prints is challenging for commercial instruments and computer vision systems.
• Existing image segmentation methods struggle with the size and complexity of colored patterns in prints.

Purpose of the Study:

• To develop and evaluate an approach for determining colorimetric attributes (L*a*b*) of multicolored materials.
• To improve the accuracy of color measurement in complex printed patterns using computer vision.

Main Methods:

• A novel approach combining global color correction and unsupervised image segmentation was developed.
• A modified Chan-Vese method was employed for segmenting printed ColorChecker charts.
• Spectrophotometric measurements and image analysis were used to extract and transform sRGB to L*a*b* values, optimizing transformation models.

Main Results:

• The modified Chan-Vese method successfully segmented multicolored printed images.
• Optimized transformation models achieved low average color differences between measured and calculated L*a*b* values.
• Mean and maximum ΔE₀₀ color differences were reported as 2.43 and 4.28, respectively, indicating satisfactory accuracy.

Conclusions:

• The proposed method offers a viable solution for accurate colorimetric analysis of multicolored prints.
• Effective image segmentation is crucial for reliable color measurement in complex printed materials.
• The approach demonstrates potential for enhancing color quality control in the printing industry.