研究目的
To develop a software program for imitating the color of an object on a PC display under different illumination conditions, converting from images taken under a white light source like D65 to simulate various light sources, aiming to accurately represent product colors in applications such as Internet sales.
研究成果
The developed software successfully imitates object colors under different illuminations by converting images from a reference light source (e.g., D65) to test illuminations. It functions well based on sample conversions, reducing the gap between standard and arbitrary illumination conditions, which is beneficial for applications like e-commerce to minimize color discrepancies in product representations. Future work could include chromatic adaptation and broader color space support.
研究不足
Chromatic adaptation was not considered in the process, which might affect color perception under different illuminations. The method relies on the assumption that spectral reflectance can be approximated with three principal components, potentially limiting accuracy for highly variable colors. The software is designed for sRGB displays and may not account for all display variations.
1：Experimental Design and Method Selection:
The study involved developing a software program in Visual C++ Ver. 6 for Windows to convert image colors based on illumination changes, using principal component analysis for spectral reflectance estimation and tristimulus value conversion. Theoretical models included formulas for color conversion and maximum luminance calculation.
2：Sample Selection and Data Sources:
Images taken under illumination similar to D65 or auto white balance set to D65 were used as original pictures. Spectral reflectance data from 1569 different colors in JIS Z8721 standard color chart were analyzed for principal components.
3：List of Experimental Equipment and Materials:
A PC with Windows OS, monitor supporting sRGB, and bitmap format images. No specific brands or models of equipment are mentioned.
4：Experimental Procedures and Operational Workflow:
Steps included: (1) loading a bitmap file, (2) converting digital pixel values to XYZ values under reference illumination (D65), (3) converting XYZ values to those under test illumination using estimated spectral reflectance, (4) converting back to digital values, and (5) displaying the converted image for visual comparison.
5：Data Analysis Methods:
Principal component analysis was used to extract characteristic vectors from spectral reflectance data. Tristimulus values were calculated and converted using matrix operations as per derived formulas.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
