# Robust deepfake detector against deep image watermarking

## TL;DR

This paper introduces a deepfake detection model that performs well even when images have deep watermarks, improving accuracy significantly over existing methods.

## Contribution

A novel multi-module deepfake detection model with Efficient Multi-scale Attention and a feature dropout module for watermark robustness.

## Key findings

- The model's accuracy remains comparable to baselines with MBRS watermarks in 50% and 100% of images.
- The model outperforms baselines by 10% and 20% in accuracy when FaceSigns watermarks are present in 50% and 100% of images.
- The model effectively eliminates redundant features to improve detection robustness against deep image watermarking.

## Abstract

Deepfake technology poses a significant threat to information security,rendering deepfake detection research crucial. However, current detection methods experience a marked performance degradation in the presence of deep watermarking within images. In this paper, we propose a multi-module model, which integrates Efficient Multi-scale Attention within Xception as the detection module and introduces a feature dropout module to eliminate redundant image features. Experimental results demonstrate that when 50% and 100% of the images in the dataset contain MBRS watermarks, the accuracy (ACC) metrics of our model are comparable to those of existingbaseline models. However, when 50% and 100% of the images contain FaceSigns watermarks, the ACC metrics of our model outperform those of other baseline models by approximately 10% and 20%, respectively.