This chapter addresses the increasing demand for realistic digital human facial appearances in the entertainment industry, such as movies and games. Existing 3D facial digitization pipelines face challenges, including labor-intensive R&D and qualitative methods for improvement. To tackle these issues, we propose optimizing the 3D face digitization pipeline using digital twins. Our approach involves using digital twins to simulate and solve digitization problems within a virtual space that accurately replicates the physical environment. By bridging the physical and virtual worlds, this method streamlines labor-intensive processes. Additionally, our system leverages quantitative evaluations of facial appearances generated in the virtual space to enhance the quality of digital human facial appearances. This approach offers a cost-effective solution for creating realistic facial appearances, making it a valuable addition to the VFX industry. In conclusion, this research introduces a promising digital twin pipeline for optimizing 3D facial appearance digitization in the VFX industry.