AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

SAFE: Semantic- and Frequency-Enhanced Curriculum for Cross-Domain Deepfake Detection

Yulin Yao, Kangfeng Zheng, Bin Wu, Chunhua Wu, Jujie Wang, Jiaqi Gao, Minjiao Yang, Dan Luo

AAAI 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Interleaved image-text generation with MLLMs · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes SAFE, a semantic- and frequency-enhanced curriculum for cross-domain deepfake detection that builds on CLIP's text-image alignment and captions.

Problem

Deepfake detectors trained on one generator family (GANs or diffusion models) degrade substantially when tested on unseen domains, and most CLIP-based detectors use CLIP only as an image-level binary classifier, ignoring its text-to-image semantic alignment.

Approach

Uses CLIP's text-image alignment with captions enriched with high-level semantics (beyond what CLIP-CAP produces), combined with frequency information in a curriculum-style training scheme. Only the abstract was available, so finer details are unknown.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add LoRA fine-tuning on CLIP visual branch (ablation)CLIP ViT-L/14 · CDFv2/DFDCP averageAUC: 0.871 -> 0.941Improved
add semantic alignment on top of LoRA (ablation)SAFE · CDFv2/DFDCP averageAUC: 0.941 -> 0.944Improved
add dual-score curriculum on top of LoRA (ablation)SAFE · CDFv2/DFDCP averageAUC: 0.941 -> 0.947Improved
combine alignment and curriculum with LoRA (ablation)SAFE · CDFv2/DFDCP averageAUC: 0.944 -> 0.956Improved
augment CLIP-CAP captions with authenticity cues (ablation)SAFE · 4-set averageAUC: 0.949 -> 0.959Improved
use raw CLIP-CAP caption instead of fake/real token (ablation)SAFE · 4-set averageAUC: 0.946 -> 0.949Improved
set loss weight lambda to 20 instead of 10 or 40 (ablation)SAFE · CDFv2/DFDCP averageAUC: 0.952 -> 0.956Improved
use convex pacing vs linear pacing (ablation)SAFE curriculum · 3 cross-manipulation sets averageAUC: 0.953 -> 0.964Improved
use convex pacing (ablation)SAFE curriculum · training timetraining time reduction: down 23.3%Improved
initial forged ratio r0=0.3 vs 0.2/0.4/0.5 (ablation)SAFE · CDFv2/SimSwap averageAUC: 0.958 -> 0.963Improved
add dual-score curriculum to ResNet-34ResNet34 · 5 cross-dataset average (frame-level)AUC: 0.767 -> 0.807Improved
add dual-score curriculum to ViT-B/16ViT-B16 · 5 cross-dataset average (frame-level)AUC: 0.845 -> 0.869Improved

Experimental setup

Builds on: CLIP CLIP-CAP

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy