Research map / 3D vision / Face and shape
FineRecon: Depth-aware Feed-forward Network for Detailed 3D Reconstruction
Noah Stier, Anurag Ranjan, Alex Colburn, Yajie Yan, Liang Yang, Fangchang Ma, Baptiste Angles
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 |
|---|---|---|---|
| supervise TSDF only at exact ground-truth points instead of interpolated voxel centers (resolution-agnostic supervision) (ablation) | FineRecon · ScanNet test | Chamfer distance: 6.40 -> 5.58 | Improved |
| add depth guidance via fused MVS TSDF channel (ablation) | FineRecon · ScanNet test | Chamfer distance: 5.58 -> 5.25 | Improved |
| add point back-projection when depth guidance is on (ablation) | FineRecon · ScanNet test | Chamfer distance: 5.25 -> 5.18 | Improved |
| add point back-projection without depth guidance (ablation) | FineRecon · ScanNet test | Chamfer distance: 5.58 -> 5.80 | Hurt |
| use TSDF fusion depth guidance instead of density volume (ablation) | FineRecon · ScanNet test | F1: 74.3 -> 75.5 | Improved |
| use TSDF fusion depth guidance instead of Gaussian weighting of features (ablation) | FineRecon · ScanNet test | F1: 73.2 -> 75.5 | Improved |
| combine TSDF fusion with Gaussian weighting (ablation) | FineRecon · ScanNet test | F1: 75.5 -> 73.8 | Hurt |
| keep image features alongside TSDF depth guidance (ablation) | FineRecon · ScanNet test | F1: 71.9 -> 75.5 | Improved |
| increase output resolution from 4 cm to 0.5 cm (ablation) | FineRecon · ScanNet test | Chamfer distance: 5.15 -> 5.19 | No clear effect |
| increase output resolution from 4 cm to 0.5 cm for SimpleRecon | SimpleRecon · ScanNet test | F1: 68.6 -> 66.9 | Hurt |
| augment predicted depth maps with random scale in [0.9, 1.1] | FineRecon · ScanNet training | overfitting to depth guidance | Improved |
Later papers that compared with this method
2 later accepted papers report FineRecon as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- AniGrad: Anisotropic Gradient-Adaptive Sampling for 3D Reconstruction From Monocular Video (CVPR 2025 · better on 3 of 5 shared results)
- DoubleTake: Geometry Guided Depth Estimation (ECCV 2024 · better on 2 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SimpleRecon — SimpleRecon: 3D Reconstruction without 3D Convolutions (ECCV 2022) · better on 4 of 4 shared results
- TransformerFusion — TransformerFusion: Monocular RGB Scene Reconstruction using Transformers (NeurIPS 2021) · 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
- Learning Neural Implicit through Volume Rendering with Attentive Depth Fusion Priors (NeurIPS 2023 · similarity 0.89)
- Atlas: End-to-End 3D Scene Reconstruction from Posed Images (ECCV 2020 · similarity 0.89)
- VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction (ICCV 2021 · similarity 0.88)
- 3D Scene Reconstruction from a Single Viewport (ECCV 2020 · similarity 0.88)
- NeuralRecon: Real-Time Coherent 3D Reconstruction From Monocular Video (CVPR 2021 · similarity 0.87)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 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.
