Research map / Image generation / Super-resolution
Channel Attention Based Iterative Residual Learning for Depth Map Super-Resolution
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use the iterative residual learning framework with channel attention, fusion, and refinement rather than DVS | Depth-map super-resolution · Interval down-sampling, ×4; Plant, 0100, Jadeplant, ls21, Room, road01 and road05 | RMSE: 0.8494 -> 0.7278; 2.8914 -> 2.7800; 2.9407 -> 2.7740; 1.8188 -> 1.6048; 1.1682 -> 1.0639; 16.0110 -> 15.6239; 26.4482 -> | Helped |
| Use the iterative residual learning framework with channel attention, fusion, and refinement rather than DVS | Depth-map super-resolution · Bi-cubic degradation with noise, ×4; Plant, 0100, Jadeplant, ls21 and Room | RMSE: 0.4565 -> 0.4004; 0.9826 -> 0.9588; 1.8458 -> 1.6765; 1.3800 -> 1.1393; 0.5903 -> 0.5351 | Helped |
| Add channel attention (ablation) | AIR · Middlebury Cones, Teddy, Tsukuba and Venus, ×4 | RMSE | Helped |
| Add more iterative sub-modules (ablation) | AIR · Middlebury Cones, Teddy, Tsukuba and Venus, ×4 bi-cubic degradation | RMSE | Helped |
| Add input loss and TGV refinement (ablation) | AIR · Middlebury Cones, Teddy, Tsukuba and Venus, ×4 bi-cubic degradation | RMSE | Helped |
| Share weights among sub-modules instead of using non-shared weights (ablation) | AIR · Interval down-sampling, ×4; Plant and Jadeplant | RMSE: 0.7278 -> 0.7300; 2.7740 -> 2.8004 | Mixed |
| Share weights among sub-modules instead of using non-shared weights (ablation) | AIR · Interval down-sampling, ×4; road01 and road05 | RMSE: 15.6239 -> 15.6305; 25.9109 -> 25.9282 | Hurt |
Closest papers by meaning
- DegBins: Degradation-Driven Binning for Depth Super-Resolution (NeurIPS 2026 · similarity 0.90)
- SGNet: Structure Guided Network via Gradient-Frequency Awareness for Depth Map Super-resolution (AAAI 2024 · similarity 0.88)
- DORNet: A Degradation Oriented and Regularized Network for Blind Depth Super-Resolution (CVPR 2025 · similarity 0.88)
- Intrinsic Phase-Preserving Networks for Depth Super Resolution (AAAI 2024 · similarity 0.87)
- Improving Depth Completion via Depth Feature Upsampling (CVPR 2024 · similarity 0.87)
- Decoupling Fine Detail and Global Geometry for Compressed Depth Map Super-Resolution (CVPR 2025 · 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.