Research map / Video / Video segmentation and tracking
Deep Optics for Video Snapshot Compressive Imaging
Ping Wang, Lishun Wang, Xin Yuan
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 |
|---|---|---|---|
| train with sensor response instead of testing pre-trained networks with it | STFormer · Six grayscale benchmarks | PSNR (dB): 36.34 -> 31.78 (mismatch) | Hurt |
| re-train with random binary mask and sensor response | STFormer · Six grayscale benchmarks | PSNR (dB): 31.78 -> 27.66 (re-trained) | Hurt |
| train under deep optics framework (learned structural mask + sensor response) | STFormer · Six grayscale benchmarks | PSNR (dB): 36.34 -> 36.69 | Improved |
| train under deep optics framework | E2E-CNN · Six grayscale benchmarks | PSNR (dB): 29.45 -> 32.42 | Improved |
| train under deep optics framework | RevSCI · Six grayscale benchmarks | Q-Score: 51.21 -> 52.74 | Improved |
| learn structural mask rather than random (ablation) | Res2former (4-bit) · Six grayscale benchmarks | PSNR (dB): 34.95 -> 35.98 | Improved |
| increase learned mask bit depth from 1 to 4 (ablation) | Res2former · Six grayscale benchmarks | PSNR (dB): 34.62 -> 35.98 | Improved |
| increase bit depth 2 to 3 for random structural mask (ablation) | Res2former · Six grayscale benchmarks | PSNR (dB): 35.03 -> 34.91 | No clear effect |
| scale up Res2former to Res2former-L | Res2former · Six grayscale benchmarks | PSNR (dB): 35.98 -> 36.56 | Improved |
| replace STFormer with Res2former (temporal-only attention) | decoder · Six grayscale benchmarks | PSNR (dB): 36.69 -> 35.98 | Mixed |
Closest papers by meaning
- Spatial-Temporal Transformer for Video Snapshot Compressive Imaging (TPAMI 2022 · similarity 0.91)
- Memory-Efficient Network for Large-Scale Video Compressive Sensing (CVPR 2021 · similarity 0.91)
- MetaSCI: Scalable and Adaptive Reconstruction for Video Compressive Sensing (CVPR 2021 · similarity 0.90)
- Plug-and-Play Algorithms for Video Snapshot Compressive Imaging (TPAMI 2021 · similarity 0.90)
- Untrained Neural Nets for Snapshot Compressive Imaging: Theory and Algorithms (NeurIPS 2024 · similarity 0.89)
- SCINeRF: Neural Radiance Fields from a Snapshot Compressive Image (CVPR 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.
