Look Locally Infer Globally: A Generalizable Face Anti-Spoofing Approach

TL;DR

This paper introduces SSR-FCN, a local, self-supervised face anti-spoofing method that generalizes well to unseen spoof types while being computationally efficient and interpretable.

Contribution

It proposes a novel self-supervised regional learning framework for face anti-spoofing that enhances generalization to unknown spoof types.

Findings

Achieves 65% TDR @ 2% FDR on diverse spoof types

Maintains high efficiency (< 4 ms) on GPU

Performs competitively on standard benchmarks

Abstract

State-of-the-art spoof detection methods tend to overfit to the spoof types seen during training and fail to generalize to unknown spoof types. Given that face anti-spoofing is inherently a local task, we propose a face anti-spoofing framework, namely Self-Supervised Regional Fully Convolutional Network (SSR-FCN), that is trained to learn local discriminative cues from a face image in a self-supervised manner. The proposed framework improves generalizability while maintaining the computational efficiency of holistic face anti-spoofing approaches (< 4 ms on a Nvidia GTX 1080Ti GPU). The proposed method is interpretable since it localizes which parts of the face are labeled as spoofs. Experimental results show that SSR-FCN can achieve TDR = 65% @ 2.0% FDR when evaluated on a dataset comprising of 13 different spoof types under unknown attacks while achieving competitive performances under standard benchmark datasets (Oulu-NPU, CASIA-MFSD, and Replay-Attack).