Research map / 3D vision / Depth and stereo
Domain Reduction Strategy for Non-Line-of-Sight Imaging
Hyunbo Shim, In Cho, Daekyu Kwon, Seon Joo 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 |
|---|---|---|---|
| add domain reduction with coarse-to-fine strategy (ablation) | optimization-based NLOS reconstruction · Confocal 32x32 Bunny | optimization time for 1000 iterations: 1087 s -> 54 s | Improved |
| add domain reduction without coarse-to-fine (ablation) | optimization-based NLOS reconstruction · Confocal 32x32 Bunny | optimization time for 1000 iterations: 1087 s -> 134 s | Improved |
| add domain reduction at high resolution (ablation) | NLOS reconstruction · N = 128 hidden volume | GPU memory (MB): 4065 -> 1659 | Improved |
| add domain reduction at low resolution (ablation) | NLOS reconstruction · N = 32 hidden volume | GPU memory (MB): 1187 -> 1211 | No clear effect |
| add jointly optimized noise parameters on top of L1 regularization (ablation) | NLOS reconstruction · Stanford Statue and Dragon | visual artifacts (qualitative) | Improved |
| model cosine factor and distance fall-off in light propagation (ablation) | NLOS reconstruction · ZNLOS Bunny | depth RMSE (qualitative) | Improved |
| use continuous random point sampling instead of fixed voxel-center sampling (ablation) | NLOS reconstruction · ZNLOS and real-world | reconstruction accuracy (qualitative) | Improved |
| increase scan resolution above 32x32 (ablation) | Ours · ZNLOS Bunny | reconstruction quality (qualitative) | No clear effect |
| optimization method versus FFT-based DLCT on sparse scans | Ours · ZNLOS Bunny confocal 32x32 | normal MAE (5%): 0.3796 -> 0.1147 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Gram — GRAM: Generative Radiance Manifolds for 3D-Aware Image Generation (CVPR 2022) · better on 2 of 2 shared results
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
- Beyond Volumetric Albedo -- A Surface Optimization Framework for Non-Line-Of-Sight Imaging (CVPR 2019 · similarity 0.91)
- Non-Line-of-Sight Imaging With Signal Superresolution Network (CVPR 2023 · similarity 0.90)
- Non-Line-of-Sight Surface Reconstruction Using the Directional Light-Cone Transform (CVPR 2020 · similarity 0.90)
- Enhancing Non-line-of-sight Imaging via Learnable Inverse Kernel and Attention Mechanisms (ICCV 2023 · similarity 0.89)
- Dual-branch Graph Feature Learning for NLOS Imaging (AAAI 2025 · similarity 0.88)
- Virtual Scanning: Unsupervised Non-line-of-sight Imaging from Irregularly Undersampled Transients (NeurIPS 2024 · similarity 0.88)
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.
