Research map / 3D vision / Depth and stereo
A Lightweight Approach for On-The-Fly Reflectance Estimation
Kihwan Kim, Jinwei Gu, Stephen Tyree, Pavlo Molchanov, Matthias Niessner, Jan Kautz
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 |
|---|---|---|---|
| add the Ec regularization to Grouplet with RMSE1 loss (ablation) | Grouplet · SynBRDF test | RMSE: 0.432 -> 0.455 | Hurt |
| add the Ec regularization to Grouplet with RMSE1 loss (ablation) | Grouplet · SynBRDF user study | user rank: 2 -> 1 | Improved |
| add the Ec regularization to HemiCNN with CubeRoot loss (ablation) | HemiCNN · SynBRDF test | RMSE: 0.439 -> 0.583 | Hurt |
| add the Ec regularization to HemiCNN with CubeRoot loss (ablation) | HemiCNN · SynBRDF user study | user rank: 4 -> 6 | Hurt |
| add the Ec regularization when training on real RGBD scans (ablation) | HemiCNN and Grouplet · real data (KinectFusion scans) | qualitative plausibility | Improved |
| switch HemiCNN loss from RMSE1 to RMSE2 (Lab-based representation) (ablation) | HemiCNN · SynBRDF test | RMSE: 0.419 -> 0.564 | Hurt |
| switch Grouplet loss from RMSE1 to RMSE2 (ablation) | Grouplet · SynBRDF user study | user rank: 2 -> 7 | Hurt |
| choice of distance loss Ed (RMSE1, RMSE2, CubeRoot) (ablation) | HemiCNN and Grouplet · SynBRDF test | RMSE and gloss accuracy | Mixed |
| increase number of observations per voxel (ablation) | HemiCNN, Grouplet-fast, Grouplet-slow · SynBRDF test | error: saturates around 30 | Improved |
| use more voxel samples (Grouplet-slow, 354 voxels, instead of fast, 20) (ablation) | Grouplet · SynBRDF | inference time: 5ms -> 90ms | Hurt |
| use Grouplet instead of HemiCNN | BRDF estimator · SynBRDF | gloss vs diffuse albedo estimation | Mixed |
Closest papers by meaning
- Shape and Spatially-Varying Reflectance Estimation from Virtual Exemplars (TPAMI 2016 · similarity 0.90)
- Deep 3D Capture: Geometry and Reflectance From Sparse Multi-View Images (CVPR 2020 · similarity 0.90)
- Reflectance Capture Using Univariate Sampling of BRDFs (ICCV 2017 · similarity 0.90)
- Shape and Reflectance Estimation in the Wild (TPAMI 2016 · similarity 0.89)
- SurfaceNet: Adversarial SVBRDF Estimation From a Single Image (ICCV 2021 · similarity 0.89)
- Real-Time Neural BRDF with Spherically Distributed Primitives (CVPR 2024 · similarity 0.87)
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.
