Research map / Image generation / Super-resolution
Zero-Shot Dual-Lens Super-Resolution
Read the paperCheck 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 |
|---|---|---|---|
| Adding degradation-invariant alignment losses (spatial adversarial, frequency adversarial, feature contrastive) to L2 alignment and degradation-aware training (ablation) | ZeDuSR · HCI new 2x IG | PSNR: 30.48 to 31.01 dB | Helped |
| Degradation-aware training strategy (DaTS) for patch cropping (ablation) | ZeDuSR · HCI new 2x IG | PSNR: 30.89 to 31.01 dB | Helped |
| Initialize SR network from bicubic-pretrained model (ZeDuSR*) | ZeDuSR · HCI new 2x IG | PSNR: 31.01 to 31.17 dB | Helped |
| Swap alignment backbone (FlowNet-S to SPyNet or PWCNet) (ablation) | ZeDuSR · HCI new 2x IG/AG | PSNR: IG 31.05 (SPyNet), 29.98 (PWCNet) vs 31.01; AG 29.97 and 30.06 vs 30.02 | No clear effect |
| Swap SR backbone (RCAN to CSNLN or SwinIR) within ZeDuSR (ablation) | ZeDuSR · HCI new 2x IG | PSNR: 31.03 (CSNLN), 31.04 (SwinIR) vs 31.01 (RCAN) | No clear effect |
| Dual-lens zero-shot ZeDuSR vs. dual-lens/RefSR/SISR baselines | ZeDuSR · Synthesized IG/AG 2x and 4x (Table 1) | PSNR: superior in most cases | Helped |
| Blind SISR (DAN, DCLS) vs ZeDuSR on AG-kernel-pool training | ZeDuSR · HCI new and Middlebury2021 (Table 2) | PSNR: DCLS 30.63 beats ZeDuSR 31.01 on HCI new IG SSIM 0.8651 vs 0.8529; ZeDuSR better overall | Mixed |
Closest papers by meaning
- KeDuSR: Real-World Dual-Lens Super-Resolution via Kernel-Free Matching (AAAI 2024 · similarity 0.89)
- Continuous Optical Zooming: A Benchmark for Arbitrary-Scale Image Super-Resolution in Real World (CVPR 2024 · similarity 0.86)
- PocketSR: The Super-Resolution Expert in Your Pocket Mobiles (NeurIPS 2025 · similarity 0.84)
- Edge-SD-SR: Low Latency and Parameter Efficient On-device Super-Resolution with Stable Diffusion via Bidirectional Conditioning (CVPR 2025 · similarity 0.84)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.84)
- Towards Realistic Data Generation for Real-World Super-Resolution (ICLR 2025 · similarity 0.84)
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.