Research map / Image generation / Super-resolution
Perceptual Deep Depth Super-Resolution
Oleg Voynov, Alexey Artemov, Vage Egiazarian, Alexander Notchenko, Gleb Bobrovskikh, Evgeny Burnaev, Denis Zorin
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 depth-difference loss with visual difference loss (MSG to MSG-V) | MSG · SimGeo average x4 | RMSEv (x1e-3): 194 -> 99 | Improved |
| replace depth-difference loss with visual difference loss (MSG to MSG-V) | MSG · SimGeo average x4 | DSSIMv (x1e-3): 374 -> 95 | Improved |
| replace depth-difference loss with visual difference loss (MSG to MSG-V) | MSG · SimGeo average x4 | LPIPSv (x1e-3): 569 -> 194 | Improved |
| replace depth-difference loss with visual difference loss (MSG to MSG-V) | MSG · SimGeo average x4 | RMSEd (mm): 39 -> 96 | Hurt |
| replace depth-difference loss with visual difference loss (DIP to DIP-V) | DIP · SimGeo average x4 | DSSIMv (x1e-3): 887 -> 313 | Improved |
| replace depth-difference loss with visual difference loss (DIP to DIP-V) | DIP · SimGeo average x4 | RMSEv (x1e-3): 395 -> 147 | Improved |
| replace depth-difference loss with visual difference loss (DIP to DIP-V) | DIP · SimGeo average x4 | RMSEd (mm): 52 -> 33 | Improved |
| use RMSEv rather than DSSIM as the training loss (ablation) | depth super-resolution optimization · single example (Figure 3) | DSSIM: 0.396 -> 0.254 | Improved |
| evaluate with RMSEd versus perceptual metrics (correlation) | quality measures · super-resolution results (Figure 4) | Pearson correlation with LPIPS: 0.85 | Mixed |
Closest papers by meaning
- Photometric Depth Super-Resolution (TPAMI 2019 · similarity 0.86)
- Neural RGB(r)D Sensing: Depth and Uncertainty From a Video Camera (CVPR 2019 · similarity 0.86)
- Deep Depth Completion of a Single RGB-D Image (CVPR 2018 · similarity 0.85)
- A Two-Streamed Network for Estimating Fine-Scaled Depth Maps From Single RGB Images (ICCV 2017 · similarity 0.85)
- Single-Image Depth Perception in the Wild (NeurIPS 2016 · similarity 0.85)
- Depth Prompting for Sensor-Agnostic Depth Estimation (CVPR 2024 · similarity 0.85)
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.
