Research map / 3D vision / Depth and stereo
Curvature Regularization for Non-Line-of-Sight Imaging From Under-Sampled Data
Rui Ding, Juntian Ye, Qifeng Gao, Feihu Xu, Yuping Duan
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 signal-domain curvature regularization (dual-domain model) | curvNLOS object-domain · Bowling (64x64) | Accuracy: 0.9558 -> 0.9585 | Improved |
| add signal-domain curvature regularization (dual-domain model) | curvNLOS object-domain · Bowling (8x8) | Accuracy: 0.9236 -> 0.9431 | Improved |
| add signal-domain curvature regularization (dual-domain model) | curvNLOS object-domain · Bowling (6x6) | Accuracy: 0.9226 -> 0.9358 | Improved |
| add signal-domain curvature regularization (dual-domain model) | curvNLOS object-domain · Bowling (4x4) | Accuracy: 0.8782 -> 0.8977 | Improved |
| replace SPIRAL sparsity prior with object-domain curvature regularization | SPIRAL · Bowling (4x4) | PSNR: 12.4577 -> 14.5064 | Improved |
| dual-domain model instead of object-domain | curvNLOS · Bowling (64x64) | PSNR: 16.2712 -> 16.1188 | Hurt |
| use TSC curvature function instead of TRV (ablation) | Algorithm 1 · Bowling (6x6) | Accuracy: 0.9177 -> 0.9226 | Mixed |
| use TSC curvature function instead of TAC (ablation) | Algorithm 2 · Bowling (6x6) | Accuracy: 0.9290 -> 0.9358 | Improved |
| use TSC curvature function instead of TRV (ablation) | Algorithm 2 · Bowling (6x6) | PSNR: 14.8599 -> 14.7758 | Hurt |
| add smoothing post-processing (ablation) | SPIRAL · under-sampled Bowling | visual quality | Improved |
| reduce scan points from 24x24 to 16x16 | curvNLOS · Teaser | total time (s): 429.50 -> 218.34 | Improved |
| use curvNLOS instead of SPIRAL | iterative reconstruction · Teaser (16x16) | reconstruction time (s): 444.24 -> 21.34 | Improved |
Closest papers by meaning
- Deep Non-line-of-sight Imaging from Under-scanning Measurements (NeurIPS 2023 · similarity 0.87)
- Non-Line-of-Sight Imaging With Signal Superresolution Network (CVPR 2023 · similarity 0.87)
- Domain Reduction Strategy for Non-Line-of-Sight Imaging (ECCV 2024 · similarity 0.87)
- Virtual Scanning: Unsupervised Non-line-of-sight Imaging from Irregularly Undersampled Transients (NeurIPS 2024 · similarity 0.86)
- Adaptive 3D Reconstruction via Diffusion Priors and Forward Curvature-Matching Likelihood Updates (NeurIPS 2025 · similarity 0.84)
- Dual-branch Graph Feature Learning for NLOS Imaging (AAAI 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.
