Research map / Image generation / Image restoration
BNUDC: A Two-Branched Deep Neural Network for Restoring Images From Under-Display Cameras
Jaihyun Koh, Jangho Lee, Sungroh Yoon
Check 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 |
|---|---|---|---|
| separate high- and low-frequency restoration into two branches (ablation) | BNUDC · POLED test set | PSNR: 32.68 -> 33.19 | Improved |
| add a 1D affine transform in the HFR branch (ablation) | BNUDC single-branch configuration · POLED test set | PSNR: 32.68 -> 32.70 | Improved |
| add a 1D affine transform in the LFR branch (ablation) | BNUDC dual-branch configuration · POLED test set | PSNR: 33.19 -> 33.27 | Improved |
| use a 3D rather than 1D affine transform in the LFR branch (ablation) | BNUDC dual-branch configuration · POLED test set | PSNR: 33.27 -> 33.32 | Improved |
| apply inverse color filtering to training images (ablation) | BNUDC dual-branch, 3D-affine configuration · POLED test set | PSNR: 33.32 -> 33.39 | Improved |
| use a 1D instead of 3D affine transform in the LFR branch (ablation) | BNUDC · TOLED test set | PSNR: 38.22 -> 38.26 | Improved |
| use a 1D instead of 3D affine transform in the LFR branch (ablation) | BNUDC · TOLED test set | SSIM: 0.9798 -> 0.9800 | Mixed |
| use a 1D instead of 3D affine transform in the LFR branch (ablation) | BNUDC · TOLED test set | DISTS: 0.0964 -> 0.0942 | Improved |
| restore images without an external PSF prior | BNUDC · SYNTH test set | PSNR: 43.27 -> 45.78 | Improved |
Later papers that compared with this method
4 later accepted papers report BNUDC as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- UCMNet: Uncertainty-Aware Context Memory Network for Under-Display Camera Image Restoration (CVPR 2026 · better on 9 of 9 shared results)
- FSI: Frequency and Spatial Interactive Learning for Image Restoration in Under-Display Cameras (ICCV 2023 · better on 4 of 4 shared results)
- Under-Display Camera Image Restoration with Scattering Effect (ICCV 2023 · better on 3 of 3 shared results)
- Generating Aligned Pseudo-Supervision From Non-Aligned Data for Image Restoration in Under-Display Camera (CVPR 2023 · better on 1 of 1 shared result)
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
- Panel-Specific Degradation Representation for Raw Under-Display Camera Image Restoration (ECCV 2024 · similarity 0.92)
- Decoupling Degradations with Recurrent Network for Video Restoration in Under-Display Camera (AAAI 2024 · similarity 0.91)
- Image Restoration for Under-Display Camera (CVPR 2021 · similarity 0.89)
- Under-Display Camera Image Restoration with Scattering Effect (ICCV 2023 · similarity 0.89)
- FSI: Frequency and Spatial Interactive Learning for Image Restoration in Under-Display Cameras (ICCV 2023 · similarity 0.88)
- DDR: Exploiting Deep Degradation Response as Flexible Image Descriptor (NeurIPS 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.
