Research map / Architectures and efficiency / Attention and transformers
Enhancing Non-line-of-sight Imaging via Learnable Inverse Kernel and Attention Mechanisms
Yanhua Yu, Siyuan Shen, Zi Wang, Binbin Huang, Yuehan Wang, Xingyue Peng, Suan Xia, Ping Liu, Ruiqian Li, Shiying Li
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 |
|---|---|---|---|
| learn the inverse kernel with attention guidance instead of using physics-based kernels | NLOS imaging network · Poses dataset | PSNR: 21.43 -> 24.59 (vs LFE) | Improved |
| apply learnable kernel with attention for classification | NLOS digit classification · Digits dataset | accuracy: 0.89 -> 0.94 | Improved |
| use low-frequency-guided cross-attention (LSHC) for the high-frequency kernel instead of self-attention only (LS, LSHS) (ablation) | learnable inverse kernel · Poses dataset | qualitative reconstruction | Improved |
| change the low/high-frequency kernel split from 50-50 to 25-75 or 75-25 (ablation) | learnable inverse kernel · Motorbikes dataset | PSNR: 26.61 -> 26.59 (25-75), 26.61 (75-25) | No clear effect |
| train on one dataset and test on another with the learned kernel | learnable inverse kernel · cross-dataset (Motorbikes and Poses) | qualitative reconstruction | Improved |
Closest papers by meaning
- Physics to the Rescue: Deep Non-Line-of-Sight Reconstruction for High-Speed Imaging (TPAMI 2022 · similarity 0.91)
- NLOST: Non-Line-of-Sight Imaging With Transformer (CVPR 2023 · similarity 0.90)
- Deep Non-line-of-sight Imaging from Under-scanning Measurements (NeurIPS 2023 · similarity 0.89)
- Virtual Scanning: Unsupervised Non-line-of-sight Imaging from Irregularly Undersampled Transients (NeurIPS 2024 · similarity 0.87)
- Dual-branch Graph Feature Learning for NLOS Imaging (AAAI 2025 · similarity 0.87)
- NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems (NeurIPS 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.
