Research map / 3D vision / Depth and stereo
Differentiable Diffusion for Dense Depth Estimation From Multi-View Images
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Optimize noisy sparse points by differentiable diffusion against multi-view reprojection | Sparse-point depth diffusion · Piano-MVS | MSE: 2.56 -> 0.81 | Helped |
| Optimize noisy sparse points by differentiable diffusion against multi-view reprojection | Sparse-point depth diffusion · Living Room-MVS | MSE: 1.37 -> 0.30 | Helped |
| Initialize with dense-COLMAP-derived labels at RGB edges (ablation) | Ours · Living Room-MVS | MSE: 0.30 -> 0.17 | Helped |
| Initialize with dense-COLMAP-derived labels at RGB edges (ablation) | Ours · Piano-MVS | MSE: 0.81 -> 0.69 | Helped |
| Use a weighted median post-processing filter (ablation) | Ours · Dino | MSE * 100: 0.86 -> 0.54 | Helped |
| Use a weighted median post-processing filter (ablation) | Ours · Dino | BP(0.01): 25.6 -> 39.6 | Hurt |
| Use VGG16 feature distance instead of RGB distance | Self-supervised reprojection loss · Light-field depth estimation | warping loss performance | Helped |
| Increase the initial point set from approximately 700 to approximately 2,800 points (ablation) | Ours · Multi-view stereo | MSE: 0.30 -> 0.17; 0.81 -> 0.69 | Helped |
| Optimize points to reduce diffusion errors from noisy or spurious samples | Ours · Synthetic light fields | BP(0.1) | Mixed |
Closest papers by meaning
- Zero-Shot Novel View and Depth Synthesis with Multi-View Geometric Diffusion (CVPR 2025 · similarity 0.88)
- Multi-view Reconstruction via SfM-guided Monocular Depth Estimation (CVPR 2025 · similarity 0.88)
- DepthFM: Fast Generative Monocular Depth Estimation with Flow Matching (AAAI 2025 · similarity 0.88)
- Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation (CVPR 2024 · similarity 0.88)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.88)
- Jasmine: Harnessing Diffusion Prior for Self-supervised Depth Estimation (NeurIPS 2025 · similarity 0.88)
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.