Research map / 3D vision / Depth and stereo
G2-MonoDepth: A General Framework of Generalized Depth Inference From Monocular RGB+X Data
Haotian Wang, Meng Yang, Nanning Zheng
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 |
|---|---|---|---|
| use ReZero blocks instead of BN (ablation) | G2-MonoDepth · KITTI synthetic (0.1% sparsity) | RMSE: 16.46 -> 12.03 | Improved |
| use ReZero blocks instead of BN (ablation) | G2-MonoDepth · NYUv2 synthetic (0.1% sparsity) | RMSE: 6.49 -> 6.02 | Improved |
| use Sobel operator instead of differential operator (ablation) | G2-MonoDepth · KITTI synthetic (0.1% sparsity) | RMSE: 29.66 -> 31.32 | Hurt |
| use Sobel operator instead of differential operator (ablation) | G2-MonoDepth · KITTI synthetic (10% sparsity) | RMSE: 5.46 -> 5.18 | Improved |
| use mean/mean deviation (G2S) instead of median standardization (MS) (ablation) | G2-MonoDepth loss · KITTI synthetic (100% sparsity) | RMSE: 18.98 -> 7.49 | Improved |
| use mean/mean deviation (G2S) instead of median standardization (MS) (ablation) | G2-MonoDepth loss · NYUv2 synthetic (10% sparsity) | RMSE: 15.75 -> 4.28 | Improved |
| use mean/mean deviation (G2S) instead of z-score (ZS) standardization (ablation) | G2-MonoDepth loss · NYUv2 synthetic (10% sparsity) | RMSE: 4.02 -> 4.28 | Hurt |
| use mean/mean deviation (G2S) instead of z-score (ZS) standardization (ablation) | G2-MonoDepth loss · Ibims synthetic (0% sparsity) | SRMSE: 0.92 -> 0.83 | Improved |
| use G2-MonoDepth over MiDas with fewer training data (0.23 million vs 19 million) and 18M vs 105M parameters (ablation) | MiDas · Depth estimation, real-world (6 datasets) | average rank on SRMSE: 2.3 -> 1.7 | Improved |
| add sparse depth input (0% to 0.1% valid pixels) | G2-MonoDepth · synthetic data | RMSE | Improved |
| use unified G2-MonoDepth instead of JIFF | JIFF · NYUv2 depth enhancement (real) | RMSE: 3.06 -> 3.87 | Hurt |
| use G2-MonoDepth instead of GuideNet | GuideNet · KITTI depth completion 64-line (real) | RMSE: 3.17 -> 4.23 | Hurt |
Later papers that compared with this method
3 later accepted papers report G2-MonoDepth as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- Large Depth Completion Model from Sparse Observations (ICLR 2026 · better on 2 of 2 shared results)
- PacGDC: Label-Efficient Generalizable Depth Completion with Projection Ambiguity and Consistency (ICCV 2025 · better on 4 of 4 shared results)
- OMNI-DC: Highly Robust Depth Completion with Multiresolution Depth Integration (ICCV 2025 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- JIFF — JIFF: Jointly-Aligned Implicit Face Function for High Quality Single View Clothed Human Reconstruction (CVPR 2022) · better on 1 of 2 shared results
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
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.88)
- Vision-Language Embodiment for Monocular Depth Estimation (CVPR 2025 · similarity 0.88)
- SQLdepth: Generalizable Self-Supervised Fine-Structured Monocular Depth Estimation (AAAI 2024 · similarity 0.88)
- GeoDepth: From Point-to-Depth to Plane-to-Depth Modeling for Self-Supervised Monocular Depth Estimation (CVPR 2025 · similarity 0.88)
- UniDepth: Universal Monocular Metric Depth Estimation (CVPR 2024 · similarity 0.88)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 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.
