Research map / 3D vision / Face and shape
Directed Ray Distance Functions for 3D Scene Reconstruction
Nilesh Kulkarni, Justin Johnson, David F. Fouhey
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 |
|---|---|---|---|
| predict ray distance instead of scene distance (UDF to URDF) | distance-function predictor · Matterport3D scene | F1: 64.7 -> 68.7 | Improved |
| add direction (sign) to the ray distance (URDF to DRDF) (ablation) | ray distance function · Matterport3D scene | F1: 68.7 -> 71.9 | Improved |
| add direction to the ray distance (URDF to DRDF) (ablation) | ray distance function · Matterport3D occluded rays | F1: 20.7 -> 27.3 | Improved |
| add direction to the ray distance (URDF to DRDF) (ablation) | ray distance function · 3DFront occluded rays | F1: 30.2 -> 52.6 | Improved |
| add direction to the ray distance (URDF to DRDF) (ablation) | ray distance function · ScanNet scene | F1: 58.2 -> 60.9 | Improved |
| use DRDF instead of ray occupancy (ORF) | ray predictor · 3DFront scene | F1: 59.6 -> 63.4 | Improved |
| use DRDF instead of LDI | layered depth predictor · Matterport3D scene | F1: 67.4 -> 71.9 | Improved |
| add confidence channels to LDI | LDI · Matterport3D scene | F1: 67.4 -> 57.7 | Hurt |
| use DRDF instead of sign-agnostic learning (SAL) | ray predictor · Matterport3D scene | F1: 34.3 -> 71.9 | Improved |
| use DRDF instead of SAL on ScanNet | ray predictor · ScanNet scene | F1: 57.7 -> 60.9 | Mixed |
| use local-minima decoding instead of absolute thresholding for UDF (ablation) | UDF · Matterport3D scene | F1: 14.7 on F1 | Improved |
| use NMS-with-threshold decoding instead of zero-crossings of gradient for URDF (ablation) | URDF · Matterport3D scene | F1: 11 on F1 | Improved |
Closest papers by meaning
- Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set (NeurIPS 2024 · similarity 0.86)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 2024 · similarity 0.86)
- Sensing Surface Patches in Volume Rendering for Inferring Signed Distance Functions (AAAI 2025 · similarity 0.86)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · similarity 0.86)
- RaySt3R: Predicting Novel Depth Maps for Zero-Shot Object Completion (NeurIPS 2025 · similarity 0.85)
- 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 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.
