Research map / Image generation / Image restoration
Matching in the Dark: A Dataset for Matching Image Pairs of Low-Light Scenes
Wenzheng Song, Masanori Suganuma, Xing Liu, Noriyuki Shimobayashi, Daisuke Maruta, Takayuki Okatani
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 |
|---|---|---|---|
| use direct RAW input with SID enhancement instead of standard RIP (ablation) | SuperPoint + SuperGlue · MID indoor | mean N_tau at 5 degrees: 0.219 -> 0.619 | Improved |
| use SID instead of Direct-HistEq on RAW input (ablation) | SuperPoint + SuperGlue · MID indoor | mean N_tau at 5 degrees: 0.548 -> 0.619 | Improved |
| add BM3D denoising to direct RAW histogram equalization (ablation) | SuperPoint · MID indoor | mean N_tau at 5 degrees: 0.548 -> 0.596 | Improved |
| add SuperGlue matching in place of nearest-neighbor search (ablation) | SuperPoint with SID · MID outdoor | mean N_tau at 5 degrees: 0.500 -> 0.575 | Improved |
| add SuperGlue matching in place of nearest-neighbor search (ablation) | SuperPoint with standard RIP · MID indoor | mean N_tau at 5 degrees: 0.223 -> 0.219 | No clear effect |
| use SuperPoint instead of RootSIFT (ablation) | SID enhancer · MID indoor | mean N_tau at 5 degrees: 0.369 -> 0.583 | Improved |
| replace histogram equalization with MIRNet after standard RIP (ablation) | SuperPoint · MID indoor | mean N_tau at 5 degrees: 0.421 -> 0.381 | Hurt |
| replace histogram equalization with CLAHE after standard RIP (ablation) | SuperPoint · MID indoor | mean N_tau at 5 degrees: 0.421 -> 0.275 | Hurt |
| use NG-RANSAC instead of RANSAC (ablation) | RootSIFT with SID · MID indoor | mean N_tau at 5 degrees: 0.369 -> 0.363 | No clear effect |
| use SID instead of BM3D denoising on RAW input (ablation) | SuperPoint · MID indoor | mean N_tau at 5 degrees: 0.596 -> 0.583 | No clear effect |
| manually choose the 14-bit RAW range instead of automatic enhancers (ablation) | SuperPoint · MID | mean N_tau at 5 degrees | Improved |
Closest papers by meaning
- DarkFeat: Noise-Robust Feature Detector and Descriptor for Extremely Low-Light RAW Images (AAAI 2023 · similarity 0.87)
- Multi-Object Tracking in the Dark (CVPR 2024 · similarity 0.85)
- Towards Robust Event-guided Low-Light Image Enhancement: A Large-Scale Real-World Event-Image Dataset and Novel Approach (CVPR 2024 · similarity 0.84)
- XFeat: Accelerated Features for Lightweight Image Matching (CVPR 2024 · similarity 0.84)
- OmniGlue: Generalizable Feature Matching with Foundation Model Guidance (CVPR 2024 · similarity 0.84)
- DarkIR: Robust Low-Light Image Restoration (CVPR 2025 · similarity 0.83)
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.
