Atlases Are Already Inside: Recovering Population Templates from Pretrained Diffusion Models

2026-09-24

AI Summary

This study addresses the challenge of extracting population templates from pretrained diffusion models without retraining. To this end, it proposes an inference-time sampler that leverages the implicit structural representations encoded within the model to directly generate intrinsic atlases. This work provides the first demonstration that atlas construction can emerge as a byproduct of generative modeling, enabling zero-retraining, registration-free template recovery across domains while supporting multimodal and subgroup analyses. Evaluated as a registration target across multiple datasets, the proposed approach achieves state-of-the-art or competitive performance and successfully recapitulates established signatures of healthy aging.

Abstract

We present a new inference-time sampler for diffusion models that gives a pretrained model a capability it was never trained for: constructing the atlas of the population it synthesizes. The sampler converges from every random seed to the population's central anatomy, which we call the \emph{intrinsic atlas}. The advantage is threefold. (1) It requires no retraining. A diffusion model that has already learned a coherent population, including the released ones, yields its atlas in a single inference pass without involving deformable registration. (2) It applies to multiple domains, such as brain MRI, chest X-ray, faces, and 3D shapes. (3) It extends to subpopulations. One age-conditioned model gives an atlas at any age in its training range, and the resulting family reproduces the CSF expansion of healthy aging. Evaluated as a registration target, the intrinsic atlas is best or second-best on every dataset against classical and learned templates, and the most central template on held-out brain MRI cohorts. Atlas construction can be reframed as a byproduct of generative modeling: a diffusion model is a learned representation of population structure, and the atlas is what it already contains.

Problem

• atlas construction
• population template
• diffusion models
• generative modeling

Innovation

• Diffusion Models
• Atlas Construction
• Inference-time Sampler
• Intrinsic Atlas
• Generative Modeling