Research map / Video / Video segmentation and tracking
Deep Appearance Maps
Maxim Maximov, Laura Leal-Taixe, Mario Fritz, Tobias Ritschel
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 |
|---|---|---|---|
| represent appearance as a 4D deep appearance map instead of multi-view reflectance map RM++ | DAM · Synthetic point light, novel view | DSSIM: .181 -> .144 | Improved |
| represent appearance as a 4D deep appearance map instead of multi-view reflectance map RM++ | DAM · Synthetic environment map, novel view | DSSIM: .193 -> .164 | Improved |
| represent appearance as a 4D deep appearance map instead of RM++ | DAM · Synthetic point light, same view | DSSIM: .143 -> .105 | Improved |
| represent appearance as a 4D deep appearance map instead of RM++ | DAM · Synthetic environment map, same view | DSSIM: .160 -> .123 | Improved |
| replace per-material SGD fitting with a learning-to-learn network predicting DAM weights | DAM · Synthetic point light, novel view | DSSIM: .144 -> .165 | Hurt |
| replace per-material SGD fitting with a learning-to-learn network predicting DAM weights | DAM · Synthetic environment map, novel view | DSSIM: .164 -> .173 | Hurt |
| replace per-material SGD fitting with a learning-to-learn network | DAM · Synthetic environment map, same view | DSSIM: .123 -> .131 | Hurt |
| replace per-material SGD fitting with a learning-to-learn network | DAM · Synthetic point light, same view | DSSIM: .105 -> .106 | No clear effect |
| use joint estimation-and-segmentation instead of k-means segmentation then per-segment estimation | DAM · MultiMaterial point light, novel view | DSSIM: .172 -> .161 | Improved |
| use joint estimation-and-segmentation instead of k-means segmentation | DAM · MultiMaterial environment map, novel view | DSSIM: .164 -> .154 | Improved |
| use joint estimation-and-segmentation instead of k-means segmentation | DAM · MultiMaterial point light, same view | DSSIM: .132 -> .113 | Improved |
| use joint estimation-and-segmentation instead of k-means segmentation | DAM · MultiMaterial environment map, same view | DSSIM: .136 -> .122 | Improved |
Closest papers by meaning
- Deep Reflectance Maps (CVPR 2016 · similarity 0.90)
- Reflectance and Natural Illumination from Single-Material Specular Objects Using Deep Learning (TPAMI 2017 · similarity 0.89)
- Diffusion Reflectance Map: Single-Image Stochastic Inverse Rendering of Illumination and Reflectance (CVPR 2024 · similarity 0.87)
- Deep 3D Capture: Geometry and Reflectance From Sparse Multi-View Images (CVPR 2020 · similarity 0.86)
- DeepSurfels: Learning Online Appearance Fusion (CVPR 2021 · similarity 0.86)
- 3D Appearance Super-Resolution With Deep Learning (CVPR 2019 · 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.
