AI Paper Atlas

Research map / Detection and segmentation / CNN features

Beyond [CLS] Token: Query-Driven Token-Level Forgery Purification for Generalizable Deepfake Detection

Changshuo Wang, Jiangming Wang, Ke-Yue Zhang, Taiping Yao, Shouhong Ding, Shunli Wang, Ran Yi, Lizhuang Ma

CVPR 2026 · Detection and segmentation › CNN features · approach: Forgery detection

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace [CLS] token with learnable query tokens (ablation)Effort r=1 · CDF-v1/FSh/DFDCP avgaverage AUC: 0.914 -> 0.925Improved
add fake-likelihood contrastive learning (FLCL) to query tokens (ablation)QTFP · CDF-v1/FSh/DFDCP avgaverage AUC: 0.925 -> 0.931Improved
add real-attention alignment loss (ablation)QTFP · CDF-v1/FSh/DFDCP avgaverage AUC: 0.931 -> 0.939Improved
apply FLCL to [CLS] token instead of query tokens (ablation)Effort r=1 · CDF-v1/FSh/DFDCP avgaverage AUC: 0.914 -> 0.929 (vs 0.931 with queries)Mixed
add QTFP to full fine-tuning (ablation)CLIP FFT · DFDC/DFDCP/WDF avgaverage AUC: 0.868 -> 0.895Improved
add QTFP to LoRA r=4 (ablation)CLIP LoRA · DFDC/DFDCP/WDF avgaverage AUC: 0.877 -> 0.898Improved
add QTFP to Effort r=1 (ablation)CLIP Effort · DFDC/DFDCP/WDF avgaverage AUC: 0.867 -> 0.903Improved
add QTFP to BEiTv2 backbone (ablation)BEiTv2 LoRA r=4 · CDF-v2/FSh avgaverage AUC: 0.723 -> 0.781Improved
add QTFP to DINOv3 backbone (ablation)DINOv3 LoRA r=4 · CDF-v2/FSh avgaverage AUC: 0.854 -> 0.881Improved
add QTFP to Siglip2 backbone (ablation)Siglip2 LoRA r=4 · CDF-v2/FSh avgaverage AUC: 0.867 -> 0.879Improved
add QTFP to PE backbone (ablation)PE LoRA r=4 · CDF-v2/FSh avgaverage AUC: 0.908 -> 0.926Improved
add QTFP to CLIP backbone (ablation)CLIP LoRA r=4 · CDF-v2/FSh avgaverage AUC: 0.909 -> 0.932Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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