Research map / Image generation / Autoregressive and latent generation
Anatomica: Localized Control over Geometric and Topological Properties for Anatomical Diffusion Models
Karim Kadry, Abdalla Abdelwahed, Ajay Manicka, Naravich Chutisilp, Farhad R. Nezami, Elazer R. Edelman
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 |
|---|---|---|---|
| guide unconditional diffusion with topological potentials | Anatomica-L · Atrial Separation | B0 precision (%): 7.81 -> 78.9 | Improved |
| guide unconditional diffusion with topological potentials | Anatomica-L · Calcium Count | B0 precision (%): 0.00 -> 60.9 | Improved |
| guide unconditional diffusion with topological potentials | Anatomica-L · Vertebrae Connectivity | B1 precision (%): 8.59 -> 26.6 | Improved |
| guide unconditional diffusion with topological potentials | Anatomica-L · Vertebrae Connectivity | B2 precision (%): 12.5 -> 7.03 | Hurt |
| guide unconditional diffusion with topological potentials | Anatomica-L · Atrial Separation | 1-NNA: 0.578 -> 0.606 | Hurt |
| use inference-time geometric guidance instead of explicit conditioning | Anatomica-L · Right Ventricle | Mass error (x1e5): 154.5 -> 17.5 | Improved |
| use inference-time geometric guidance instead of implicit conditioning | Anatomica-L · Mitral Valve | FMD: 314.8 -> 88.8 | Improved |
| use guidance instead of implicit conditioning | Anatomica-L · Aortic Trunk | Mass error (x1e5): 0.81 -> 2.38 | Hurt |
| use neural field decoder with L-parsing instead of convolutional decoder with V-parsing (ablation) | Anatomica-L · Mitral Valve | Centroid error (x1e4): 41.4 -> 11.4 | Improved |
| use neural field decoder with L-parsing instead of convolutional decoder with V-parsing (ablation) | Anatomica-L · Right Ventricle | Mass error (x1e5): 12.3 -> 17.5 | Hurt |
| reduce local L-parsing resolution from high to low (ablation) | Anatomica-L · Right Ventricle | relative speed: 2.48 -> 10.40 | Improved |
| reduce local L-parsing resolution from high to low (ablation) | Anatomica-L · Right Ventricle | FMD: 91.16 -> 105.57 | Hurt |
Closest papers by meaning
- CardioComposer: Leveraging Differentiable Geometry for Compositional Control of Anatomical Diffusion Models (ICLR 2026 · similarity 0.88)
- AnaDiffusion: Anatomically Compositional Latent Diffusion for Controllable 3D Brain MRI Generation (NeurIPS 2026 · similarity 0.87)
- Conformable Convolution for Topologically Constrained Learning of Complex Anatomical Structures (AAAI 2026 · similarity 0.86)
- Parametric Surface Diffeomorphometry for Low Dimensional Embeddings of Dense Segmentations and Imagery (TPAMI 2016 · similarity 0.85)
- TopoDiffusionNet: A Topology-aware Diffusion Model (ICLR 2025 · similarity 0.84)
- MultiMorph: On-demand Atlas Construction (CVPR 2025 · similarity 0.83)
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.
