Research map / Vision-language and multimodal / Cross-modal learning
Detecting and Grounding Multi-Modal Media Manipulation and Beyond
Rui Shao, Tianxing Wu, Jianlong Wu, Liqiang Nie, 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 global manipulation-aware contrastive loss (MAC-G) (ablation) | HAMMER · DGM4 | binary AUC: 92.77 -> 93.19 | Improved |
| add multi-label classification loss (ablation) | HAMMER · DGM4 | multi-label mAP: 27.22 -> 86.22 | Improved |
| add image manipulation grounding loss (ablation) | HAMMER · DGM4 | image grounding IoUmean: 4.69 -> 76.45 | Improved |
| add text manipulation grounding loss (ablation) | HAMMER · DGM4 | text grounding F1: 22.53 -> 71.35 | Improved |
| add grounding and contrastive losses to binary-classification-only training (ablation) | HAMMER · DGM4 | binary ACC: 83.81 -> 86.39 | Improved |
| add local-view manipulation-aware contrastive loss (MAC-L) (ablation) | HAMMER · DGM4 | text grounding F1: 71.35 -> 72.59 | Improved |
| use manipulation-aware contrastive learning instead of normal contrastive learning (ablation) | HAMMER · DGM4 | binary AUC: 92.95 -> 93.19 | Improved |
| add momentum models (ablation) | HAMMER · DGM4 | binary AUC: 92.96 -> 93.19 | Improved |
| add momentum models (ablation) | HAMMER · DGM4 | text grounding precision: 76.13 -> 75.01 | Hurt |
| add text modality to image-only model (ablation) | HAMMER · DGM4 text-forgery split | text grounding F1: 37.77 -> 72.08 | Improved |
| add text modality to image-only model (ablation) | HAMMER · DGM4 image-forgery split | binary AUC: 93.96 -> 94.40 | Improved |
| replace [CLS] token with local patch attentional aggregation (LPAA) for bbox grounding (ablation) | HAMMER · DGM4 | image grounding | Improved |
Closest papers by meaning
- Unleashing the Potential of Consistency Learning for Detecting and Grounding Multi-Modal Media Manipulation (CVPR 2025 · similarity 0.92)
- IDseq: Decoupled and Sequentially Detecting and Grounding Multi-Modal Media Manipulation (AAAI 2025 · similarity 0.88)
- Text-Guided Fine-grained Counterfactual Inference for Short Video Fake News Detection (AAAI 2025 · similarity 0.86)
- DAMMFND: Domain-Aware Multimodal Multi-view Fake News Detection (AAAI 2025 · similarity 0.85)
- FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Models (ICLR 2025 · similarity 0.85)
- ForgerySleuth: Empowering Multimodal Large Language Models for Image Manipulation Detection (NeurIPS 2025 · similarity 0.85)
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.
