Research map / 3D vision / Depth and stereo
Masked Depth Modeling for Spatial Perception
Bin Tan, CHANGJIANG SUN, Xiage Qin, Hanat Adai, Zelin Fu, Tianxiang Zhou, Han Zhang, YINGHAO XU, Xing Zhu, Yujun Shen, Nan Xue
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 MDM-pretrained encoder instead of DINOv2 | MoGe · NYUv2 monocular depth | Aff-inv depth REL: 0.056 -> 0.044 | Improved |
| use MDM-pretrained encoder instead of DINOv2 | MoGe · ETH3D monocular depth | Scl-inv depth REL: 0.137 -> 0.091 | Improved |
| use MDM-pretrained encoder instead of DINOv2 | MoGe · KITTI monocular depth | Scl-inv point REL: 0.204 -> 0.208 | Mixed |
| use MDM-pretrained encoder instead of DINOv2 | MoGe · DDAD monocular depth | Scl-inv depth REL: 0.230 -> 0.230 | No clear effect |
| use MDM-pretrained encoder instead of DINOv2 | MoGe · Sintel monocular depth | Scl-inv depth REL: 0.318 -> 0.305 | Improved |
| use LingBot-Depth prior instead of vanilla FoundationStereo backbone | FoundationStereo · HAMMER stereo, epoch 5 | EPE: 0.46 -> 0.27 | Improved |
| use LingBot-Depth prior instead of vanilla FoundationStereo backbone | FoundationStereo · Booster stereo, epoch 5 | EPE: 1.00 -> 0.86 | Improved |
| use DepthAnythingV2 prior via MoGe instead of LingBot-Depth prior | FoundationStereo · HAMMER stereo, epoch 5 | EPE: 0.27 -> 2.53 | Hurt |
| use MDM pretrained model instead of PromptDA | depth completion · DIODE-Indoor extreme masking | RMSE: 0.465 -> 0.221 | Improved |
| use MDM pretrained model instead of best baseline | depth completion · ETH3D sparse SfM (outdoor) | MAE: 0.215 -> 0.223 | Mixed |
| use LingBot-Depth refined depth instead of raw depth | dexterous grasping · Toy Car grasping | successful grasps of 20: 9 -> 16 | Improved |
| use LingBot-Depth refined depth instead of raw depth | dexterous grasping · Transparent storage box grasping | successful grasps of 20: N/A -> 10 | Improved |
Closest papers by meaning
- DepthCues: Evaluating Monocular Depth Perception in Large Vision Models (CVPR 2025 · similarity 0.87)
- SharpDepth: Sharpening Metric Depth Predictions Using Diffusion Distillation (CVPR 2025 · similarity 0.86)
- Depth Pro: Sharp Monocular Metric Depth in Less Than a Second (ICLR 2025 · similarity 0.86)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.86)
- Vision-Language Embodiment for Monocular Depth Estimation (CVPR 2025 · similarity 0.86)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.86)
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.
