Research map / 3D vision / Depth and stereo
Single-Shot Hyperspectral-Depth Imaging With Learned Diffractive Optics
Seung-Hwan Baek, Hayato Ikoma, Daniel S. Jeon, Yuqi Li, Wolfgang Heidrich, Gordon Wetzstein, Min H. Kim
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 |
|---|---|---|---|
| replace Fresnel DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | PSNR (dB): 27.96 -> 29.31 | Improved |
| replace Spiral DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | PSNR (dB): 26.90 -> 29.31 | Improved |
| replace Fisher DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | PSNR (dB): 28.51 -> 29.31 | Improved |
| replace Fisher DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | SSIM: 0.79 -> 0.81 | Improved |
| replace Fresnel DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | depth MAE (m): 0.15 -> 0.12 | Improved |
| replace Fisher DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | depth RMSE (m): 0.23 -> 0.20 | Improved |
| replace Spiral DOE with end-to-end learned DOE (ablation) | HS-D reconstruction network · HS-D dataset | depth RMSE (m): 0.32 -> 0.20 | Improved |
| replace light-field compressive HS-D method (Feng et al.) with learned DOE (ablation) | single-shot HS-D imaging · HS-D dataset | PSNR (dB): 23.62 -> 29.31 | Improved |
| replace light-field compressive HS-D method (Feng et al.) with learned DOE (ablation) | single-shot HS-D imaging · HS-D dataset | depth RMSE (m): 0.57 -> 0.20 | Improved |
| replace light-field compressive HS-D method (Feng et al.) with learned DOE (ablation) | single-shot HS-D imaging · HS-D dataset | SSIM: 0.76 -> 0.81 | Improved |
| replace light-field compressive HS-D method (Feng et al.) with learned DOE (ablation) | single-shot HS-D imaging · HS-D dataset | depth MAE (m): 0.30 -> 0.12 | Improved |
| use Fisher-information DOE as initialization instead of Fresnel (CRLB discriminating power) (ablation) | DOE design · CRLB analysis | CRLB: 2.73 -> 1.36 | Improved |
Closest papers by meaning
- Quantization-Aware Deep Optics for Diffractive Snapshot Hyperspectral Imaging (CVPR 2022 · similarity 0.87)
- Filterless Snapshot Hyperspectral Imaging using Guided Patch Diffusion (ECCV 2026 · similarity 0.87)
- HIR-Diff: Unsupervised Hyperspectral Image Restoration Via Improved Diffusion Models (CVPR 2024 · similarity 0.85)
- Dense Dispersed Structured Light for Hyperspectral 3D Imaging of Dynamic Scenes (CVPR 2025 · similarity 0.85)
- Dispersed Structured Light for Hyperspectral 3D Imaging (CVPR 2024 · similarity 0.84)
- SPECAT: SPatial-spEctral Cumulative-Attention Transformer for High-Resolution Hyperspectral Image Reconstruction (CVPR 2024 · 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.
