Research map / 3D vision / Depth and stereo
FCFR-Net: Feature Fusion based Coarse-to-Fine Residual Learning for Depth Completion
Lina Liu, Xibin Song, Xiaoyang Lyu, Junwei Diao, Mengmeng Wang, Yong Liu, Liangjun Zhang
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 sparse depth input with dense coarse depth input (ablation) | STD depth completion framework · KITTI validation (selected subset) | RMSE (mm): 942.65 -> 845.55 | Improved |
| add residual learning instead of directly predicting depth (ablation) | dense-input STD framework · KITTI validation (selected subset) | RMSE (mm): 845.55 -> 816.94 | Improved |
| use two feature encoders instead of one (ablation) | residual depth completion framework · KITTI validation (selected subset) | RMSE (mm): 816.94 -> 812.92 | Improved |
| add channel shuffle between color and depth features (ablation) | two-encoder residual framework · KITTI validation (selected subset) | RMSE (mm): 812.92 -> 805.23 | Improved |
| use energy-based fusion instead of concatenation (ablation) | channel-shuffle residual framework · KITTI validation (selected subset) | RMSE (mm): 805.23 -> 802.62 | Improved |
| use energy-based fusion instead of addition (ablation) | channel-shuffle residual framework · KITTI validation (selected subset) | RMSE (mm): 806.26 -> 802.62 | Improved |
| use FCFR-Net rather than STD | KITTI depth completion · KITTI test | RMSE (mm) | Improved |
| use FCFR-Net rather than PwP | NYUDv2 depth completion with 500 sampled points · NYUDv2 | RMSE (m): 0.112 -> 0.106 | Improved |
Later papers that compared with this method
1 later accepted paper reports FCFRNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- RigNet: Repetitive Image Guided Network for Depth Completion (ECCV 2022 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CSPN++: Learning Context and Resource Aware Convolutional Spatial Propagation Networks for Depth Completion (AAAI 2020 · similarity 0.89)
- RGBD Based Dimensional Decomposition Residual Network for 3D Semantic Scene Completion (CVPR 2019 · similarity 0.88)
- RigNet: Repetitive Image Guided Network for Depth Completion (ECCV 2022 · similarity 0.88)
- Completion as Enhancement: A Degradation-Aware Selective Image Guided Network for Depth Completion (CVPR 2025 · similarity 0.88)
- Bilateral Propagation Network for Depth Completion (CVPR 2024 · similarity 0.87)
- Improving Depth Completion via Depth Feature Upsampling (CVPR 2024 · similarity 0.87)
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.
