AtlasMorph: Learning conditional deformable templates for brain MRI

2025-11-17

AI Summary

In medical image analysis, computationally efficient fixed, unconditional templates fail to capture population-level anatomical variability, degrading registration and segmentation performance. To address this, we propose the first end-to-end learnable conditional deformable brain template generation framework, where subject-specific attributes (e.g., age, sex) serve as conditioning inputs, enabling joint optimization of template generation and image registration. Our method employs a 3D convolutional registration network augmented with segmentation label supervision to dynamically synthesize population-specific templates. Evaluated on multi-center brain MRI datasets, our approach significantly outperforms conventional unconditional and unsupervised templates, achieving measurable gains in registration accuracy (Dice ↑3.2%, HD95 ↓18.7%) and population representativeness. This work establishes a novel, interpretable, and generalizable paradigm for anatomical prior modeling tailored to personalized imaging analysis.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Structured Learning
• Humans and AI: Brain-Sensing and Analysis

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Deformable templates, or atlases, are images that represent a prototypical anatomy for a population, and are often enhanced with probabilistic anatomical label maps. They are commonly used in medical image analysis for population studies and computational anatomy tasks such as registration and segmentation. Because developing a template is a computationally expensive process, relatively few templates are available. As a result, analysis is often conducted with sub-optimal templates that are not truly representative of the study population, especially when there are large variations within this population. We propose a machine learning framework that uses convolutional registration neural networks to efficiently learn a function that outputs templates conditioned on subject-specific attributes, such as age and sex. We also leverage segmentations, when available, to produce anatomical segmentation maps for the resulting templates. The learned network can also be used to register subject images to the templates. We demonstrate our method on a compilation of 3D brain MRI datasets, and show that it can learn high-quality templates that are representative of populations. We find that annotated conditional templates enable better registration than their unlabeled unconditional counterparts, and outperform other templates construction methods.

Problem

• Creating population-representative brain MRI templates efficiently
• Addressing sub-optimal templates in neuroimaging analysis
• Generating conditional templates using demographic attributes

Innovation

• Learning conditional deformable templates using neural networks
• Generating templates based on age and sex attributes
• Producing anatomical segmentation maps from available segmentations