计算机视觉
人工智能
计算机科学
繁殖
图像(数学)
生物
生态学
作者
Zhengnan Ye,Haisong Xu,Jiaci Deng,Jueqin Qiu,Yiming Huang,Li Li
摘要
An end-to-end image reproduction method is proposed to render cultural heritage images under arbitrary illuminating conditions. When displayed on a characterized display, the appearance of paper-based cultural heritages is reproduced as actually observed under the target illumination, by leveraging spectral images of the artworks and spectral power distribution (SPD) of the identical illumination. Psychophysical experiments are conducted to verify that the proposed method outperforms the traditional pipelines on reproducing real scenes, including the CIECAM02 color appearance model. The method has the potential to be applied in the digitization and exhibition of paper cultural heritages.
科研通智能强力驱动
Strongly Powered by AbleSci AI