The project, led by Prof. Wu Hongzhi from CCST, was supported by the NSFC key Program in 2023. The project plans to establish a high-fidelity model optimization framework that can accurately approximate the collected data and ensure reconstruction quality under non-collection conditions, propose a joint expression of differentiable geometry and materials, develop a high-precision reconstruction method for complex appearances, and develop high-performance active acquisition The equipment breaks through the quality bottleneck of cultural relic digitization and achieves photo-level reconstruction of the appearance of cultural relics under any lighting and viewing angle (SSIM>0.95/PSNR>30dB). This project is of great significance to improving the research and application level of high-end cultural relics digital equipment in our country.
