ReGenMorph: Visibly Realistic GAN Generated Face Morphing Attacks by Attack Re-generation

Face morphing attacks aim at creating face images that are verifiable to be the of multiple identities, which can lead building faulty identity links in operations like border checks. While a morphed detector (MFD), training on all possible attack types is essential achieve good detection performance. Therefore, investigating new methods drives generalizability MADs. Creating was performed image level, by landmark interpolation, or latent-space manipulating latent vectors generative adversarial network. The earlier results varying blending artifacts and latter synthetic-like striping artifacts. This work presents novel pipeline, ReGenMorph, eliminate LMA using GAN-based generation, as well as, manipulation space, resulting visibly realistic compared previous works. generated ReGenMorph appearance recent approaches evaluated for recognition vulnerability detectability, whether known unknown attacks.