Research map / 3D vision / Depth and stereo
Deep RGB-D Canonical Correlation Analysis For Sparse Depth Completion
Read the paperCheck 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 CFCNet instead of the strongest listed baseline | NYUv2, stereo sparsifier, 100 points · NYUv2, stereo sparsifier, 100 points | RMSE: 0.629 -> 0.580 | Helped |
| Use CFCNet instead of the strongest listed baseline | NYUv2, stereo sparsifier, 200 points · NYUv2, stereo sparsifier, 200 points | RMSE: 0.603 -> 0.531 | Helped |
| Use CFCNet instead of the strongest listed baseline | NYUv2, stereo sparsifier, 500 points · NYUv2, stereo sparsifier, 500 points | RMSE: 0.529 -> 0.476 | Helped |
| Use CFCNet instead of Zhou et al. | Cityscape-to-KITTI evaluation, 100 points · Cityscape to KITTI, 100 points | RMSE: 7.580 -> 5.827 | Helped |
| Use CFCNet instead of CNN-SLAM | SLAM RGBD cross-dataset evaluation · SLAM RGBD, average | percentage of correctly estimated depth: 22.46 -> 57.81 | Helped |
| Add smoothness loss (ablation) | CFCNet · KITTI odometry, 500 points | RMSE: 2.967 -> 2.964 | Helped |
| Use complementary RGB rather than dense RGB with sparse depth (ablation) | CFCNet · NYUv2, stereo sparsifier, 100 points | RMSE: 0.638 -> 0.631 | Helped |
| Use complementary RGB rather than dense RGB with sparse depth (ablation) | CFCNet · NYUv2, stereo sparsifier, 100 points | MAE: 0.479 -> 0.473 | Helped |
| Add sparse depth to dense RGB input (ablation) | CFCNet · NYUv2, stereo sparsifier, 100 points | RMSE: 0.740 -> 0.638 | Helped |
| Increase the number of stereo-sampled depth points (ablation) | CFCNet · NYUv2, stereo sparsifier | RMSE: 0.715 -> 0.231 | Helped |
| Use dense RGB instead of complementary RGB (ablation) | CFCNet with sparse depth · NYUv2, stereo sparsifier, 100 points | delta_3: 98.1 -> 97.7 | Hurt |
Closest papers by meaning
- Completion as Enhancement: A Degradation-Aware Selective Image Guided Network for Depth Completion (CVPR 2025 · similarity 0.88)
- Efficient Indoor Depth Completion Network Using Mask-adaptive Gated Convolution (AAAI 2025 · similarity 0.86)
- DeCoTR: Enhancing Depth Completion with 2D and 3D Attentions (CVPR 2024 · similarity 0.85)
- Improving Depth Completion via Depth Feature Upsampling (CVPR 2024 · similarity 0.85)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.85)
- Bilateral Propagation Network for Depth Completion (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.