Research map / 3D vision / Depth and stereo
GHOST: Geometry-Guided Hallucination of Opaque Surface Textures
Langxu Zhao, Zuan Gu, Tianhan Gao
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 |
|---|---|---|---|
| preprocess images with GHOST | DA3 · ClearGrasp depth | RMSE: 0.023 -> 0.018 | Improved |
| preprocess images with GHOST | MoGe2 · ClearGrasp depth | REL: 0.025 -> 0.013 | Improved |
| preprocess images with GHOST | DepthPro · ClearGrasp depth | RMSE: 0.019 -> 0.018 | Improved |
| preprocess images with GHOST | DA3 · ClearGrasp normals | mean angular error: 32.30 -> 21.26 | Improved |
| preprocess images with GHOST | VGGT · ClearPose reconstruction | F-Score: 57.28 -> 86.72 | Improved |
| preprocess images with GHOST | MapAnything · ClearPose reconstruction | F-Score: 21.94 -> 22.60 | No clear effect |
| add TransDecomp (alpha/foreground) and DAF-Net to full pipeline (ablation) | GHOST · ClearGrasp depth, MoGe2 | RMSE: 0.089 -> 0.014 | Improved |
| add DAF-Net normals to TDINO + TDecomp + GSTNet (ablation) | GHOST · ClearGrasp depth, MoGe2 | RMSE: 0.045 -> 0.014 | Improved |
| use partial GHOST pipeline instead of raw images (ablation) | GHOST ablation · ClearGrasp depth, MoGe2 | RMSE: 0.018 -> 0.089 | Hurt |
| use partial GHOST configurations (ablation) | GHOST ablation · ClearPose reconstruction, pi3 | F-Score: 61.61 -> 26.35 / 31.36 / 39.83 | Hurt |
| use full GHOST on pi3 | GHOST · ClearPose reconstruction, pi3 | F-Score: 61.61 -> 75.26 | Improved |
| combine TransDINO with DINOv3 features instead of baselines on ClearGrasp | TransDINO · ClearGrasp segmentation | mIoU: 71.40 -> 92.44 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 2 of 2 shared results
- MoGe2 — MoGe-2: Accurate Monocular Geometry with Metric Scale and Sharp Details (NeurIPS 2025) · better on 2 of 2 shared results
- DUSt3R — DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024) · 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
- From Transparent to Opaque: Rethinking Neural Implicit Surfaces with $\alpha$-NeuS (NeurIPS 2024 · similarity 0.87)
- 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects (CVPR 2026 · similarity 0.87)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 2024 · similarity 0.87)
- Fix False Transparency by Noise Guided Splatting (NeurIPS 2025 · similarity 0.86)
- Geometry Field Splatting with Gaussian Surfels (CVPR 2025 · similarity 0.85)
- DarkVGGT: Seeing Through Darkness Using Thermal Geometry without Daylight Tax (NeurIPS 2026 · 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.
