Big Brother is Watching: Proactive Deepfake Detection via Learnable Hidden Face
Hongbo Li, Shangchao Yang, Ruiyang Xia, Lin Yuan, Xinbo Gao
TL;DR
The paper addresses the limited generalization of passive deepfake detectors by proposing a proactive defense that embeds a learnable face inside a host face using a semi-fragile invertible steganography network. A learnable secret template is optimized to resemble a neutral appearance, and a simulated transmission channel with benign and malicious manipulations enables robust, cross-type deepfake detection guided by a multi-term loss $L_{Total} = \lambda_{1} L_{Steg} + \lambda_{2} L_{Sec} + \lambda_{3} L_{Rec} + \lambda_{4} L_{Det}$. The framework combines an INN-based steganography module, a channel simulator, secret image restoration, and a patchGAN-based detector to reliably distinguish manipulated from benign-transformed images. Extensive cross-dataset experiments show superiority over both passive and proactive baselines in detection accuracy and steganographic image quality, demonstrating practical robustness for real-world deployment.
Abstract
As deepfake technologies continue to advance, passive detection methods struggle to generalize with various forgery manipulations and datasets. Proactive defense techniques have been actively studied with the primary aim of preventing deepfake operation effectively working. In this paper, we aim to bridge the gap between passive detection and proactive defense, and seek to solve the detection problem utilizing a proactive methodology. Inspired by several watermarking-based forensic methods, we explore a novel detection framework based on the concept of ``hiding a learnable face within a face''. Specifically, relying on a semi-fragile invertible steganography network, a secret template image is embedded into a host image imperceptibly, acting as an indicator monitoring for any malicious image forgery when being restored by the inverse steganography process. Instead of being manually specified, the secret template is optimized during training to resemble a neutral facial appearance, just like a ``big brother'' hidden in the image to be protected. By incorporating a self-blending mechanism and robustness learning strategy with a simulative transmission channel, a robust detector is built to accurately distinguish if the steganographic image is maliciously tampered or benignly processed. Finally, extensive experiments conducted on multiple datasets demonstrate the superiority of the proposed approach over competing passive and proactive detection methods.