Research map / 3D vision / Face and shape
Totems: Physical Objects for Verifying Visual Integrity
Jingwei Ma, Lucy Chai, Minyoung Huh, Tongzhou Wang, Ser-Nam Lim, Phillip Isola, Antonio Torralba
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 joint totem position optimization (ablation) | totem reconstruction · Real images | Reconstruction L1: 0.149 -> 0.109 | Improved |
| add joint totem position optimization (ablation) | totem reconstruction · Synthetic Env map scene | Reconstruction L1: 0.173 -> 0.103 | Improved |
| add joint totem position optimization (ablation) | totem reconstruction · Synthetic Box scene | Reconstruction L1: 0.057 -> 0.054 | Improved |
| add joint totem position optimization (ablation) | totem reconstruction · Synthetic Box scene | Pose L1: 0.008 -> 0.108 | Hurt |
| add joint totem position optimization (ablation) | totem reconstruction · Synthetic Env map scene | Pose L1: 0.060 -> 0.027 | Improved |
| use ground-truth oracle totem positions instead of estimated (ablation) | totem reconstruction · Synthetic Env map scene | Reconstruction L1: 0.103 -> 0.040 | Improved |
| add totem optimization to detector (L1 score) (ablation) | totem detection · CAF | patch-wise mAP: 0.485 -> 0.554 | Improved |
| add totem optimization to detector (LPIPS score) (ablation) | totem detection · Splice | patch-wise mAP: 0.449 -> 0.739 | Improved |
| add totem optimization to detector (LPIPS score) (ablation) | totem detection · Color | patch-wise mAP: 0.954 -> 0.946 | No clear effect |
| use totem-based method instead of ManTra-Net | totem detection with LPIPS · CAF | patch-wise mAP: 0.181 -> 0.666 | Improved |
| use LPIPS instead of L1 as patch score (ablation) | totem detection with totem opt. · Splice | patch-wise mAP: 0.638 -> 0.739 | Improved |
| increase number of totems from two to four (ablation) | totem reconstruction · Synthetic scene | reconstruction L1: 0.16 -> 0.08 | Improved |
Closest papers by meaning
- Neural Refraction Fields for Image Verification (NeurIPS 2026 · similarity 0.85)
- Focus Manipulation Detection via Photometric Histogram Analysis (CVPR 2018 · similarity 0.83)
- GlassGS: Geometry and Concept-Aware 3D Gaussian Splatting for Reflective Enclosures (ECCV 2026 · similarity 0.82)
- DepthShield: Optical Depth Encoding for Image Tampering Detection (TPAMI 2026 · similarity 0.82)
- Geometry Cloak: Preventing TGS-based 3D Reconstruction from Copyrighted Images (NeurIPS 2024 · similarity 0.81)
- Eclipse: Disambiguating Illumination and Materials using Unintended Shadows (CVPR 2024 · similarity 0.80)
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.
