Advances in Physically-Based Modeling and Inverse Design

The field of computer graphics and physics-based modeling is moving towards more realistic and accurate representations of real-world phenomena. Researchers are focusing on developing innovative methods for inverse design, physically-based modeling, and generative refinement. One of the key directions is the use of diffusion models and generative priors to advance electromagnetic inverse design and human avatar modeling. Another area of interest is the development of novel frameworks for shadow synthesis, caustic design, and surface light source representation. These advancements have the potential to enable rapid exploration of complex metasurface architectures, accelerate the development of next-generation photonic and wireless communication systems, and improve the realism and accuracy of computer-generated images. Notable papers include: PALM, a large-scale dataset for learning multi-subject hand priors, which enables realistic and relightable single-image hand avatar personalization. Diffusion-Based Electromagnetic Inverse Design of Scattering Structured Media, which presents a conditional diffusion model for electromagnetic inverse design that generates structured media geometries directly from target differential scattering cross-section profiles.

Sources

PALM: A Dataset and Baseline for Learning Multi-subject Hand Prior

Diffusion-Based Electromagnetic Inverse Design of Scattering Structured Media

Deep Inverse Shading: Consistent Albedo and Surface Detail Recovery via Generative Refinement

KPLM-STA: Physically-Accurate Shadow Synthesis for Human Relighting via Keypoint-Based Light Modeling

Computational Caustic Design for Surface Light Source

Built with on top of