Die innovative Methode GPJA ermöglicht die gemeinsame Ausrichtung von Geometrie und Photometrie für Gesichtsmeshes.
The author presents a Geometric-Photometric Joint Alignment method for aligning human expressions by combining geometry and photometric information, achieving joint alignment automatically without semantic annotation or aligned meshes. The approach utilizes differentiable rendering and multiscale regularized optimization for robust and fast convergence.