Research map / Architectures and efficiency / Attention and transformers
TransForensics: Image Forgery Localization With Dense Self-Attention
Jing Hao, Zhixin Zhang, Shicai Yang, Di Xie, Shiliang Pu
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 self-attention and positional encoding to FCN (ablation) | FCN (ResNet50) · CASIA | AUC: 0.669 -> 0.809 | Improved |
| add self-attention and positional encoding to FCN (ablation) | FCN (ResNet50) · COVERAGE | AUC: 0.715 -> 0.862 | Improved |
| add self-attention and positional encoding to FCN (ablation) | FCN (ResNet50) · IMD2020 | AUC: 0.735 -> 0.815 | Improved |
| add self-attention without positional encoding (ablation) | FCN (ResNet50) · CASIA | AUC: 0.669 -> 0.655 | Hurt |
| add positional encoding without self-attention (ablation) | FCN (ResNet50) · COVERAGE | AUC: 0.715 -> 0.707 | Hurt |
| add dense correction (ablation) | self-attention + positional encoding model · CASIA | AUC: 0.809 -> 0.837 | Improved |
| add dense correction (ablation) | self-attention + positional encoding model · COVERAGE | AUC: 0.862 -> 0.883 | Improved |
| add dense correction (ablation) | self-attention + positional encoding model · IMD2020 | AUC: 0.815 -> 0.848 | Improved |
| fuse branches by multiplication instead of addition (ablation) | TransForensics · CASIA | AUC: 0.739 -> 0.837 | Improved |
| fuse branches by multiplication instead of addition (ablation) | TransForensics · COVERAGE | AUC: 0.857 -> 0.883 | Improved |
| fuse branches by multiplication instead of addition (ablation) | TransForensics · IMD2020 | AUC: 0.827 -> 0.848 | Improved |
| use low-level C2 branch rather than deep C5 branch as output (ablation) | TransForensics · IMD2020 | AUC: 0.827 -> 0.848 | Improved |
Closest papers by meaning
- DiffForensics: Leveraging Diffusion Prior to Image Forgery Detection and Localization (CVPR 2024 · similarity 0.88)
- Learning Discriminative Noise Guidance for Image Forgery Detection and Localization (AAAI 2024 · similarity 0.86)
- FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Models (ICLR 2025 · similarity 0.86)
- A Lottery Ticket Hypothesis Approach with Sparse Fine-tuning and MAE for Image Forgery Detection and Localization (AAAI 2025 · similarity 0.84)
- MGQFormer: Mask-Guided Query-Based Transformer for Image Manipulation Localization (AAAI 2024 · similarity 0.84)
- Forgery-aware Adaptive Transformer for Generalizable Synthetic Image Detection (CVPR 2024 · similarity 0.84)
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.
