Pay Less Attention to Deceptive Artifacts: Robust Detection of Compressed Deepfakes on Online Social Networks

2025-06-25

AI Summary

Existing deepfake detection methods predominantly operate on pristine images, overlooking the JPEG compression-induced blocking artifacts ubiquitous in online social networks (OSNs). These artifacts significantly obscure forensic traces, leading to substantial performance degradation. To address this, we propose PLADA—the first deepfake detection framework that explicitly models blocking artifacts. PLADA introduces a dual-stage attention mechanism to jointly suppress compression distortions and incorporates a Block Effect Eraser and an Open Data Aggregation module, enabling robust training without large-scale paired data. By synergistically leveraging GAN- and diffusion-based architectures, PLADA extracts compression-invariant features. Extensive evaluation across 26 benchmarks demonstrates consistent superiority over state-of-the-art methods, particularly under high JPEG compression and limited paired supervision. The framework markedly enhances practical deployability in real-world OSN environments.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Learning on the Edge & Model Compression
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

With the rapid advancement of deep learning, particularly through generative adversarial networks (GANs) and diffusion models (DMs), AI-generated images, or ``deepfakes", have become nearly indistinguishable from real ones. These images are widely shared across Online Social Networks (OSNs), raising concerns about their misuse. Existing deepfake detection methods overlook the ``block effects" introduced by compression in OSNs, which obscure deepfake artifacts, and primarily focus on raw images, rarely encountered in real-world scenarios. To address these challenges, we propose PLADA (Pay Less Attention to Deceptive Artifacts), a novel framework designed to tackle the lack of paired data and the ineffective use of compressed images. PLADA consists of two core modules: Block Effect Eraser (B2E), which uses a dual-stage attention mechanism to handle block effects, and Open Data Aggregation (ODA), which processes both paired and unpaired data to improve detection. Extensive experiments across 26 datasets demonstrate that PLADA achieves a remarkable balance in deepfake detection, outperforming SoTA methods in detecting deepfakes on OSNs, even with limited paired data and compression. More importantly, this work introduces the ``block effect" as a critical factor in deepfake detection, providing a robust solution for open-world scenarios. Our code is available at https://github.com/ManyiLee/PLADA.

Problem

• Detect compressed deepfakes on social networks
• Address block effects from image compression
• Improve detection with limited paired data

Innovation

• Dual-stage attention mechanism for block effects
• Open Data Aggregation for paired and unpaired data
• Robust deepfake detection in compressed images