Research map / 3D vision / Depth and stereo
Beyond Geometry: Artistic Disparity Synthesis for Immersive 2D-to-3D
Ping Chen, Zezhou Chen, Xingpeng Zhang, Yanlin Qian, Huan Hu, Xiang Liu, Zipeng Wang, Xin Wang, Zhaoxiang Liu, Kai Wang, Shiguo Lian
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 artistic synthesis loss LArt (dual-path supervision) (ablation) | CameraNet baseline · Cinematic test set, mean over genres | std of global style s: 0.018 -> 0.009 | Improved |
| add artistic synthesis loss LArt (ablation) | CameraNet baseline · Cinematic test set, mean over genres | mean of global style s (GT: Anime 0.013, Sci-Fi 0.023, Modern 0.012): 0.030 -> 0.020 | Mixed |
| add artistic synthesis loss LArt (ablation) | CameraNet baseline · Cinematic test set, mean over genres | zero-plane shift t std: 2.35 -> 1.80 | Improved |
| add artistic synthesis loss LArt (ablation) | CameraNet baseline · Modern genre | zero-plane shift t std: 2.43 -> 2.93 | Hurt |
| add local path supervision Lpath(Mlocal) plus out-of-screen data (ablation) | Art3D · Qualitative pop-out on 2D images | pop-out effect | Improved |
| Art3D vs Depth-Anything-V2 geometric-only | Depth-Anything-V2 · User study (50 clips, 25 participants) | Overall Preference: 20.0% -> 80.0% | Improved |
| Art3D vs Depth-Anything-V2 geometric-only | Depth-Anything-V2 · User study | Visual Comfort: 41.5% -> 58.5% | Improved |
| Art3D preserves geometry in right view | Art3D · Right-view DDC-IoU | DDC-IoU: 0.85, 0.83, 0.89 on three samples | Improved |
Closest papers by meaning
- Mono2Stereo: A Benchmark and Empirical Study for Stereo Conversion (CVPR 2025 · similarity 0.86)
- PixelSynth: Generating a 3D-Consistent Experience From a Single Image (ICCV 2021 · similarity 0.86)
- A Recipe for Generating 3D Worlds from a Single Image (ICCV 2025 · similarity 0.86)
- Geometry-guided Online 3D Video Synthesis with Multi-View Temporal Consistency (CVPR 2025 · similarity 0.85)
- SimVS: Simulating World Inconsistencies for Robust View Synthesis (CVPR 2025 · similarity 0.85)
- InstaInpaint: Instant 3D-Scene Inpainting with Masked Large Reconstruction Model (NeurIPS 2025 · 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.
