Research map / Video / Video segmentation and tracking
Delving into the Local: Dynamic Inconsistency Learning for DeepFake Video Detection
Zhihao Gu, Yang Chen, Taiping Yao, Shouhong Ding, Jilin Li, Lizhuang Ma
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 |
|---|---|---|---|
| Use local snippets with Intra-SIM and Inter-SIM instead of sparse-frame temporal modeling | DeepFake video detection · FaceForensics++ LQ NT | accuracy: 0.9178 -> 0.9428 | Improved |
| Use local snippets with Intra-SIM and Inter-SIM instead of sparse-frame temporal modeling | DeepFake video detection · DFDC | accuracy: 0.8980 -> 0.9279 | Improved |
| Use local snippets with Intra-SIM and Inter-SIM instead of sparse-frame temporal modeling | DeepFake video detection · WildDeepfake | accuracy: 0.8462 -> 0.8511 | Improved |
| Use local snippets with Intra-SIM and Inter-SIM | DeepFake video detection · Celeb-DF | accuracy | No clear effect |
| Add Intra-SIM and Inter-SIM to ResNet-50 (ablation) | ResNet-50 · FaceForensics++ LQ NT | accuracy: 0.8750 -> 0.9428 | Improved |
| Add Inter-SIM only (ablation) | ResNet-50 · FaceForensics++ LQ NT | accuracy: 0.8750 -> 0.9214 | Improved |
| Add Intra-SIM only (ablation) | ResNet-50 · FaceForensics++ LQ NT | accuracy: 0.8750 -> 0.9036 | Improved |
| Add Intra-SIM and Inter-SIM together rather than either alone (ablation) | ResNet-50 · FaceForensics++ LQ NT | accuracy: 0.9214 -> 0.9428 | Improved |
| Add the bi-directional path to Inter-SIM (ablation) | Inter-SIM · FaceForensics++ LQ NT | accuracy: 0.9143 -> 0.9214 | Improved |
| Add coordinate attention to Inter-SIM with BP (ablation) | Inter-SIM · FaceForensics++ LQ NT | accuracy: 0.9214 -> 0.9286 | Improved |
| Add learnable kernels to Inter-SIM with BP and CA (ablation) | Inter-SIM · FaceForensics++ LQ NT | accuracy: 0.9286 -> 0.9428 | Improved |
| Place Intra-SIM and Inter-SIM across all backbone stages rather than a subset (ablation) | ResNet-50 · FaceForensics++ LQ NT | accuracy: 0.9357 -> 0.9428 | Improved |
Closest papers by meaning
- Generalizing Deepfake Video Detection with Plug-and-Play: Video-Level Blending and Spatiotemporal Adapter Tuning (CVPR 2025 · similarity 0.90)
- Exploiting Style Latent Flows for Generalizing Deepfake Video Detection (CVPR 2024 · similarity 0.89)
- From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes (NeurIPS 2025 · similarity 0.88)
- Towards a Universal Synthetic Video Detector: From Face or Background Manipulations to Fully AI-Generated Content (CVPR 2025 · similarity 0.88)
- DF40: Toward Next-Generation Deepfake Detection (NeurIPS 2024 · similarity 0.87)
- Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model (CVPR 2025 · similarity 0.87)
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.
