The Latent Variable Evolution process with 3D Morphable Face Model. The CMA-ES optimizer iteratively updates the albedo code a and the shape code β to generate a 3D master face that maximizes the joint false matching rate across both 2D and 3D FR systems. The FLAME model and Albedo model of a 3DMM produce a synthesized face mesh and texture, which are then rasterized and passed through the preprocessor P to create RGB-D images. These images are fed into 2D and 3D FR, where they are compared against a gallery of authentic faces 𝒯_h. The ask-and-tell loop continues for n iterations, ultimately yielding a 3D master face that simultaneously compromises both modalities.