光学
观察员(物理)
计算机科学
RGB颜色模型
人工智能
计算机视觉
物理
谱线
空间频率
一次性
原色
色差
彩色视觉
算法
杂散光
面子(社会学概念)
出处
期刊:Optics Express
[Optica Publishing Group]
日期:2026-09-17
卷期号:34 (21): 39991-39991
摘要
Observer metamerism is increasingly important in modern wide-gamut displays because narrow, saturated primaries can amplify color-matching differences among individual observers. This study combines constrained computational optimization with psychophysical white-matching experiments to design and validate RGB primary spectra that reduce observer metamerism magnitude (OMM) while maintaining specified BT.2020 gamut coverage. In the simulation, RGB primaries were modeled as Gaussian spectra, and their peak wavelengths and full widths at half maximum (FWHMs) were optimized using a genetic algorithm and individual color matching functions derived from measured observer data. As the required BT.2020 coverage increased from 85% to 95%, the optimized mean OMM increased from 2.36 to 4.15 CIEDE2000 units, revealing a clear trade-off between wide gamut and reduced observer metamerism. At 90% coverage, the optimized primary peaks were 619, 524, and 443 nm, with FWHMs of 9.9, 36.1, and 57.8 nm, respectively. Experimental validation was performed using white-matching data from 70 observers viewing 13 RGB display configurations under 2° and 10° field conditions. When the measured LED spectra were evaluated using the same model, simulated and experimental OMM values were strongly correlated, with Pearson coefficients of 0.887, 0.952, and 0.965 for the 2°, 10°, and averaged conditions, respectively. Both approaches identified very short-wavelength blue primaries as producing high OMM and indicated lower-mismatch regions near 523 nm for green and 432-449 nm for blue. These results show that the reference-anchored OMM model captures experimentally observed primary-wavelength trends and can be used to identify spectral regions predicted to reduce observer mismatch while preserving specified wide-gamut performance.
科研通智能强力驱动
Strongly Powered by AbleSci AI