Research map / Detection and segmentation / Weakly supervised detection and segmentation
Beyond Fully Supervised Pixel Annotations: Scribble-Driven Weakly-Supervised Framework for Image Manipulation Localization
Songlin Li, Guofeng Yu, Zhiqing Guo, Yunfeng Diao, Dan Ma, Gaobo Yang
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 GAFM and PFMM to baseline (ablation) | SCAF · ID average | F1: 0.533 -> 0.811 | Improved |
| add GAFM and PFMM to baseline (ablation) | SCAF · OOD average | F1: 0.283 -> 0.469 | Improved |
| add one module (GAFM per text) to baseline (ablation) | SCAF · ID average | F1: 0.533 -> 0.697 | Improved |
| add context affinity loss L_CA (ablation) | SCAF · ID average | F1: 0.627 -> 0.731 | Improved |
| add structural consistency loss L_SC (ablation) | SCAF · ID average | F1: 0.731 -> 0.755 | Improved |
| add structural consistency loss L_SC (ablation) | SCAF · OOD average | F1: 0.438 -> 0.414 | Hurt |
| add confidence-aware entropy loss L_CEM alone (ablation) | SCAF · ID average | F1: 0.731 -> 0.733 | No clear effect |
| combine L_SC and L_CEM (ablation) | SCAF · ID average | F1: 0.731 -> 0.811 | Improved |
| combine L_SC and L_CEM (ablation) | SCAF · OOD average | F1: 0.438 -> 0.469 | Improved |
| use scribble supervision instead of image-level labels | SCAF · CASIAv1 | F1: 0.349 -> 0.716 | Improved |
| use scribble supervision vs full pixel supervision (Mesorch) | SCAF · OOD average | F1: 0.261 -> 0.469 | Improved |
| use scribble supervision vs PIM on OOD | SCAF · OOD average | F1: 0.368 -> 0.469 | Improved |
Closest papers by meaning
- Training-Free Image Manipulation Localization Using Diffusion Models (AAAI 2025 · similarity 0.89)
- Can We Model the Artifacts Explicitly? Disentangle Artifacts via Pairwise Edit Relations for Image Manipulation Localization (NeurIPS 2026 · similarity 0.87)
- Scribble Hides Class: Promoting Scribble-Based Weakly-Supervised Semantic Segmentation with Its Class Label (AAAI 2024 · similarity 0.86)
- Can We Get Rid of Handcrafted Feature Extractors? SparseViT: Nonsemantics-Centered, Parameter-Efficient Image Manipulation Localization Through Spare-Coding Transformer (AAAI 2025 · similarity 0.86)
- Mesoscopic Insights: Orchestrating Multi-Scale & Hybrid Architecture for Image Manipulation Localization (AAAI 2025 · similarity 0.86)
- Less Supervision, Better Generalization: Weakly Supervised Fake Region Localization in Diffusion-Edited Images (NeurIPS 2026 · 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.
