Approximating Optimal Morphing Attacks using Template Inversion

TL;DR

This work develops a novel type of deep morphing attack based on inverting a theoretical optimal morph embedding, which is obtained as an average of the face embeddings of source images, and demonstrates that it can compete with and regularly beat the previous state of the art for deep-learning based morph generation in terms of effectiveness.

Abstract

Recent works have demonstrated the feasibility of inverting face recognition systems, enabling to recover convincing face images using only their embeddings. We leverage such template inversion models to develop a novel type ofdeep morphing attack based on inverting a theoretical optimal morph embedding, which is obtained as an average of the face embeddings of source images. We experiment with two variants of this approach: the first one exploits a fully self-contained embedding-to-image inversion model, while the second leverages the synthesis network of a pretrained StyleGAN network for increased morph realism. We generate morphing attacks from several source datasets and study the effectiveness of those attacks against several face recognition networks. We showcase that our method can compete with and regularly beat the previous state of the art for deep-learning based morph generation in terms of effectiveness, both in white-box and black-box attack scenarios, and is additionally much faster to run. We hope this might facilitate the development of large scale deep morph datasets for training detection models.

Figures (2)

• Figure 1: Illustration of the morphing process: face embeddings of the source images are extracted using the FRS $\mathcal{F}$, the corresponding optimal morph embedding is computed by interpolation in the embedding space, then fed back to the template inversion model $\mathcal{I}$ to get the morph.
• Figure 2: All types of considered deep morphs for two different pairs of source identities.