Research map / Image generation / Image restoration
An Adaptive Neural Network for Unsupervised Mosaic Consistency Analysis in Image Forensics
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Unsupervised retraining of the pretrained network directly on the test database | Proposed network · JPEG quality 95, misaligned forgeries | AUC: 0.567 to 0.746 | Helped |
| Unsupervised retraining on the test database | Proposed network · Dresden-based DB, misaligned forgeries only | AUC: 0.926 to 0.919 | No clear effect |
| Unsupervised retraining on the test database | Proposed network · Dresden-based DB, all forgeries | AUC: 0.821 to 0.817 | No clear effect |
| Proposed method evaluated on algorithms not used in pretraining | Proposed network · Unseen-algorithm subset, misaligned forgeries | AUC: 0.926 to 0.913 (pretrained) | No clear effect |
| Proposed network (pretrained) vs. VCD baseline on misaligned forgeries | Proposed network · Dresden-based DB, misaligned forgeries only | AUC: 0.926 vs 0.843 | Helped |
| Pixelwise 1x1 convolution stack and blockwise design (proposed architecture) vs. ManTraNet | Proposed network · Dresden-based DB, misaligned forgeries only | AUC: 0.926 vs 0.534 | Helped |
| Pretrained network evaluated on JPEG q95 images without adaptation | Proposed network · JPEG quality 95, misaligned forgeries | AUC: 0.926 to 0.567 (vs. unJPEG-compressed setting) | Hurt |
Closest papers by meaning
- Learning Discriminative Noise Guidance for Image Forgery Detection and Localization (AAAI 2024 · similarity 0.84)
- A Lottery Ticket Hypothesis Approach with Sparse Fine-tuning and MAE for Image Forgery Detection and Localization (AAAI 2025 · similarity 0.82)
- Forensic Self-Descriptions Are All You Need for Zero-Shot Detection, Open-Set Source Attribution, and Clustering of AI-generated Images (CVPR 2025 · similarity 0.82)
- Community Forensics: Using Thousands of Generators to Train Fake Image Detectors (CVPR 2025 · similarity 0.82)
- Reinforced Multi-teacher Knowledge Distillation for Efficient General Image Forgery Detection and Localization (AAAI 2025 · similarity 0.82)
- M²RL-Net: Multi-View and Multi-Level Relation Learning Network for Weakly-Supervised Image Forgery Detection (AAAI 2025 · similarity 0.82)
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.