Research map / 3D vision / Depth and stereo
Dynamic Fluid Surface Reconstruction Using Deep Neural Network
Simron Thapa, Nianyi Li, Jinwei Ye
Check 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 the full FSRN instead of RefractNet | Depth estimation · Synthetic fluid test set | RMSE: 0.303 -> 0.126 | Improved |
| Use the full FSRN instead of DenseDepth | Depth estimation · Synthetic fluid test set | RMSE: 0.851 -> 0.126 | Improved |
| Add the depth-normal consistency loss (ablation) | FSRN-CNN · Synthetic fluid test set, depth estimation | RMSE: 0.198 -> 0.183 | Improved |
| Add the refraction loss (ablation) | FSRN-CNN · Synthetic fluid test set, depth estimation | RMSE: 0.183 -> 0.137 | Improved |
| Add the recurrent subnet (ablation) | FSRN · Synthetic fluid test set, depth estimation | RMSE: 0.128 -> 0.126 | Improved |
| Add the depth-normal consistency loss (ablation) | FSRN-CNN · Synthetic fluid test set, normal estimation | RMSE: 0.118 -> 0.112 | Improved |
| Add the refraction loss (ablation) | FSRN-CNN · Synthetic fluid test set, normal estimation | RMSE: 0.112 -> 0.110 | Improved |
| Add the recurrent subnet (ablation) | FSRN · Synthetic fluid test set, normal estimation | RMSE: 0.098 -> 0.079 | Improved |
| Use recurrent refinement rather than framewise CNN output (ablation) | Real fluid video reconstruction · Real captured fluid videos | Temporal consistency | Improved |
Closest papers by meaning
- Stereo-Based 3D Reconstruction of Dynamic Fluid Surfaces by Global Optimization (CVPR 2017 · similarity 0.88)
- Inferring Hybrid Neural Fluid Fields from Videos (NeurIPS 2023 · similarity 0.87)
- Simulating Fluids in Real-World Still Images (ICCV 2023 · similarity 0.86)
- NeuroFluid: Fluid Dynamics Grounding with Particle-Driven Neural Radiance Fields (ICML 2022 · similarity 0.86)
- Physics Perception in Sloshing Scenes With Guaranteed Thermodynamic Consistency (TPAMI 2022 · similarity 0.86)
- Looking Into the Water by Unsupervised Learning of the Surface Shape (NeurIPS 2025 · similarity 0.84)
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.
