Deepfake Detection in Social Media: A Temporal Artifact Analysis Using 3D Convolutional Neural Networks

TL;DR

This paper introduces a 3D CNN-based deepfake detector that leverages temporal artifacts to improve detection accuracy, especially on high-quality synthetic videos, and demonstrates its effectiveness across datasets.

Contribution

The study develops a novel 3D CNN detector utilizing temporal consistency regularization, enhancing deepfake detection accuracy and generalization over existing spatial-only methods.

Findings

Achieves 92.8% accuracy on intra-dataset evaluation at 128x128 resolution.

Cross-dataset transfer accuracy reaches 76.4% without fine-tuning, improving with minimal adaptation.

Temporal artifacts provide a more robust detection signal than spatial cues, especially on high-quality fakes.

Abstract

Synthetic facial videos have proliferated across social media faster than platform moderation can respond, raising the cost of disinformation and identity-based attacks. Frame-level deepfake detectors degrade sharply as generator quality increases; high-quality 128x128 GAN output cuts spatial-only accuracy by five percentage points while leaving temporal inconsistencies largely intact. We address this gap with a 3D Convolutional Neural Network detector based on R3D-18, trained with a composite loss that combines binary cross-entropy with a temporal-consistency regularizer. The model processes 16-frame clips from the DeepfakeTIMIT dataset and is initialized from Kinetics-400 action-recognition weights. We report 92.8% accuracy on intra-dataset evaluation at 128x128 resolution; cross-dataset transfer to FaceForensics++ without fine-tuning reaches 76.4%, rising after minimal fine-tuning. Ablation studies show that transfer learning contributes 7.2 percentage points and face tracking adds 3.5 points, while temporal consistency regularization provides additional gains on high-quality fakes. The results establish that temporal artifacts generalize more broadly than spatial ones, providing a detection signal that survives social-media re-encoding.