Research map / Detection and segmentation / Open-set and few-shot detection
Robust Sequential DeepFake Detection
Rui Shao, Tianxing Wu, Ziwei Liu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add auto-regressive mechanism (ablation) | SeqFakeFormer (ResNet-50) · Seq-DeepFake components manipulation | Adap-Acc: 52.66 -> 55.30 | Improved |
| add SECA (ablation) | SeqFakeFormer (ResNet-50) · Seq-DeepFake components manipulation | Adap-Acc: 54.64 -> 55.30 | Improved |
| add SECA (ablation) | SeqFakeFormer (ResNet-34) · Seq-DeepFake attributes manipulation | Adap-Acc: 46.56 -> 48.32 | Improved |
| add auto-regressive mechanism without SECA (ablation) | SeqFakeFormer (ResNet-34) · Seq-DeepFake attributes manipulation | Fixed-Acc: 66.98 -> 66.70 | No clear effect |
| add ISM and ISC image-sequence reasoning modules (ablation) | SeqFakeFormer++ (ResNet-50) · Seq-DeepFake-P components | Adap-Acc: 44.34 -> 45.35 | Improved |
| add ISM and ISC image-sequence reasoning modules (ablation) | SeqFakeFormer++ (ResNet-34) · Seq-DeepFake-P components | Fixed-Acc: 63.51 -> 65.23 | Improved |
| add ISM and ISC (ablation) | SeqFakeFormer++ (ResNet-50) · Seq-DeepFake-P attributes | Adap-Acc: 45.43 -> 47.30 | Improved |
| add ISC only (ablation) | SeqFakeFormer++ (ResNet-50) · Seq-DeepFake-P components | Fixed-Acc: 65.19 -> 65.22 | No clear effect |
| treat as sequence prediction instead of multi-label classification (SeqFakeFormer vs DRN, MA, TS) | SeqFakeFormer (ResNet-50) · Seq-DeepFake components manipulation | Adap-Acc: 53.89 -> 55.30 | Improved |
| perturbations applied to data (Seq-DeepFake-P vs Seq-DeepFake) | SeqFakeFormer (ResNet-50) · components manipulation | Adap-Acc: 55.30 -> 44.34 | Hurt |
| use SeqFakeFormer++ instead of SeqFakeFormer on perturbed data | SeqFakeFormer (ResNet-50) · Seq-DeepFake-P attributes | Adap-Acc: 45.43 -> 47.30 | Improved |
| use SeqFakeFormer instead of DETR | DETR (ResNet-50) · Seq-DeepFake components manipulation | Adap-Acc: 49.84 -> 55.30 | Improved |
Closest papers by meaning
- DF40: Toward Next-Generation Deepfake Detection (NeurIPS 2024 · similarity 0.89)
- DiffusionFake: Enhancing Generalization in Deepfake Detection via Guided Stable Diffusion (NeurIPS 2024 · similarity 0.88)
- Standing on the Shoulders of Giants: Reprogramming Visual-Language Model for General Deepfake Detection (AAAI 2025 · similarity 0.88)
- From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes (NeurIPS 2025 · similarity 0.88)
- Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector (CVPR 2025 · similarity 0.88)
- Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model (CVPR 2025 · similarity 0.88)
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.
