MedDIFT: Multi-Scale Diffusion-Based Correspondence in 3D Medical Imaging
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918350071791616 |
|---|---|
| author | Zhang, Xingyu Reithmeir, Anna Kögl, Fryderyk Braren, Rickmer Schnabel, Julia A. Lang, Daniel M. |
| author_facet | Zhang, Xingyu Reithmeir, Anna Kögl, Fryderyk Braren, Rickmer Schnabel, Julia A. Lang, Daniel M. |
| contents | Accurate spatial correspondence between medical images is essential for longitudinal analysis, lesion tracking, and image-guided interventions. Medical image registration methods rely on local intensity-based similarity measures, which fail to capture global semantic structure and often yield mismatches in low-contrast or anatomically variable regions. Recent advances in diffusion models suggest that their intermediate representations encode rich geometric and semantic information. We present MedDIFT, a training-free 3D correspondence framework that leverages multi-scale features from a pretrained latent medical diffusion model as voxel descriptors. MedDIFT fuses diffusion activations into rich voxel-wise descriptors and matches them via cosine similarity, with an optional local-search prior. On a publicly available lung CT dataset, MedDIFT shows promising capability in identifying anatomical correspondence without requiring any task-specific model training. Ablation experiments confirm that multi-level feature fusion and modest diffusion noise improve performance. Code is available online. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_05571 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MedDIFT: Multi-Scale Diffusion-Based Correspondence in 3D Medical Imaging Zhang, Xingyu Reithmeir, Anna Kögl, Fryderyk Braren, Rickmer Schnabel, Julia A. Lang, Daniel M. Computer Vision and Pattern Recognition Accurate spatial correspondence between medical images is essential for longitudinal analysis, lesion tracking, and image-guided interventions. Medical image registration methods rely on local intensity-based similarity measures, which fail to capture global semantic structure and often yield mismatches in low-contrast or anatomically variable regions. Recent advances in diffusion models suggest that their intermediate representations encode rich geometric and semantic information. We present MedDIFT, a training-free 3D correspondence framework that leverages multi-scale features from a pretrained latent medical diffusion model as voxel descriptors. MedDIFT fuses diffusion activations into rich voxel-wise descriptors and matches them via cosine similarity, with an optional local-search prior. On a publicly available lung CT dataset, MedDIFT shows promising capability in identifying anatomical correspondence without requiring any task-specific model training. Ablation experiments confirm that multi-level feature fusion and modest diffusion noise improve performance. Code is available online. |
| title | MedDIFT: Multi-Scale Diffusion-Based Correspondence in 3D Medical Imaging |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.05571 |