Research map / Image generation / Autoregressive and latent generation
GLARE: Towards Generalizable Detection of Latent Diffusion Images with Global-Local Reconstruction Error
Jiangtao Yan, Jiazhen Ji, Zhongyu Zhang, Yuge Huang, Wenbin Wang, Shouhong Ding
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 patch-based reconstruction error to full-image error via log-ratio (ablation) | GLARE · LDMFakeDetect | AUROC: 91.54 -> 96.82 | Improved |
| add semantic complexity calibration (edge-density term) (ablation) | GLARE Delta(x) · LDMFakeDetect | AUROC: 96.82 -> 98.13 | Improved |
| use only patch-level reconstruction error (ablation) | GLARE · LDMFakeDetect | AUROC: 98.13 -> 70.11 | Hurt |
| patch size 128 instead of 32 (ablation) | GLARE · LDMFakeDetect | AUROC: 90.48 -> 98.13 | Improved |
| patch size 128 instead of 256 (ablation) | GLARE · LDMFakeDetect | AUROC: 92.02 -> 98.13 | Improved |
| use SD2-base VAE instead of other autoencoders (SDv1-1, SDv1-4, Kandinsky 2.1) (ablation) | GLARE · LDMFakeDetect | AUROC: 98.13 vs 96.97-97.11 | No clear effect |
| use GLARE instead of best training-free baseline | training-free detection · Chameleon | AUROC: 67.78 -> 82.62 | Improved |
| use GLARE instead of supervised ForgeLens | FakeInversion · FakeInversion Wurstchen2 | AUROC: 99.86 -> 88.34 | Hurt |
| apply stronger perturbations (extreme crop, blur) | GLARE · Robustness (crop, JPEG, blur, noise) | AP | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- AEROBLADE — AEROBLADE: Training-Free Detection of Latent Diffusion Images Using Autoencoder Reconstruction Error (CVPR 2024) · better on 4 of 4 shared results
- ForgeLens — ForgeLens: Data-Efficient Forgery Focus for Generalizable Forgery Image Detection (ICCV 2025) · better on 2 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- AEROBLADE: Training-Free Detection of Latent Diffusion Images Using Autoencoder Reconstruction Error (CVPR 2024 · similarity 0.90)
- Aligned Datasets Improve Detection of Latent Diffusion-Generated Images (ICLR 2025 · similarity 0.89)
- LaRE^2: Latent Reconstruction Error Based Method for Diffusion-Generated Image Detection (CVPR 2024 · similarity 0.89)
- FIRE: Robust Detection of Diffusion-Generated Images via Frequency-Guided Reconstruction Error (CVPR 2025 · similarity 0.88)
- Simpler Diffusion: 1.5 FID on ImageNet512 with Pixel-space Diffusion (CVPR 2025 · similarity 0.88)
- Shielded Diffusion: Generating Novel and Diverse Images using Sparse Repellency (ICML 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.
