Research map / Image generation / Super-resolution
AdaBridge-SR: Adaptive Bridge Matching for Real-World Image Super-Resolution
Jiangang Wang, Shangquan Sun, Aiping Zhang, Yuning Cui, Wenqi Ren
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 temporal exponent p only (ablation) | AdaBridge-SR (2-step) · DRealSR | LPIPS: 0.3122 -> 0.3375 | Hurt |
| add temporal exponent p only (ablation) | AdaBridge-SR (2-step) · DRealSR | MUSIQ: 65.6529 -> 66.5795 | Improved |
| add spatial stochasticity map M only (ablation) | AdaBridge-SR (2-step) · DRealSR | LPIPS: 0.3122 -> 0.3244 | Hurt |
| add both spatial map M and temporal exponent p (ablation) | AdaBridge-SR (2-step) · RealSR | PSNR: 24.6818 -> 24.8815 | Improved |
| add both spatial map M and temporal exponent p (ablation) | AdaBridge-SR (2-step) · DRealSR | PSNR: 27.4918 -> 27.8302 | Improved |
| add both M and p (ablation) | AdaBridge-SR (2-step) · DRealSR | LPIPS: 0.3375 -> 0.3103 | Improved |
| add stochastic noise beta 0.005 versus deterministic beta 0 (ablation) | AdaBridge-SR (4-step) · RealSR | LPIPS: 0.3284 -> 0.3138 | Improved |
| increase beta from 0.005 to 0.01 (ablation) | AdaBridge-SR (4-step) · RealSR | PSNR: 24.4671 -> 24.0997 | Hurt |
| increase beta from 0.005 to 0.01 (ablation) | AdaBridge-SR (4-step) · RealSR | MUSIQ: 70.2799 -> 70.7515 | Improved |
| increase inference steps from 1 to 4 (ablation) | AdaBridge-SR · RealSR | MUSIQ: 67.62 -> 70.28 | Improved |
| increase inference steps from 1 to 4 (ablation) | AdaBridge-SR · RealSR | LPIPS: 0.2818 -> 0.3138 | Hurt |
| increase inference steps from 1 to 4 (ablation) | AdaBridge-SR · DRealSR | LPIPS: 0.2969 -> 0.3548 | Hurt |
Closest papers by meaning
- Exploring the Design Space of Diffusion Bridge Models (NeurIPS 2025 · similarity 0.89)
- UniDB: A Unified Diffusion Bridge Framework via Stochastic Optimal Control (ICML 2025 · similarity 0.89)
- Self-Adaptive Reality-Guided Diffusion for Artifact-Free Super-Resolution (CVPR 2024 · similarity 0.87)
- DREAM: Diffusion Rectification and Estimation-Adaptive Models (CVPR 2024 · similarity 0.87)
- EditBridge: Towards Faithful and Efficient Ultra-High-Resolution Image Editing (NeurIPS 2026 · similarity 0.87)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · 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.
