An Efficient Model-Driven Groupwise Approach for Atlas Construction

2025-08-14

AI Summary

Medical image atlas construction faces dual challenges: data-driven methods suffer from poor generalizability and heavy reliance on large-scale annotated data, while model-driven approaches exhibit limited scalability and suboptimal optimization efficiency. This paper proposes DARC—a training-free, differentiable diffeomorphic model-driven group registration framework. Its core innovations include: (i) a coordinate descent optimization strategy coupled with a centrality activation function, enabling memory-efficient, one-stage registration of arbitrary-scale 3D images; and (ii) inverse deformation-based precise label propagation and novel anatomical shape synthesis. Evaluated across multiple 3D medical imaging datasets, DARC generates high-fidelity population atlases that significantly outperform state-of-the-art methods in few-shot segmentation tasks and enable high-quality anatomical shape generation. The framework achieves computational efficiency, clinical scalability, and robust performance without requiring supervised training.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• 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

Atlas construction is fundamental to medical image analysis, offering a standardized spatial reference for tasks such as population-level anatomical modeling. While data-driven registration methods have recently shown promise in pairwise settings, their reliance on large training datasets, limited generalizability, and lack of true inference phases in groupwise contexts hinder their practical use. In contrast, model-driven methods offer training-free, theoretically grounded, and data-efficient alternatives, though they often face scalability and optimization challenges when applied to large 3D datasets. In this work, we introduce DARC (Diffeomorphic Atlas Registration via Coordinate descent), a novel model-driven groupwise registration framework for atlas construction. DARC supports a broad range of image dissimilarity metrics and efficiently handles arbitrary numbers of 3D images without incurring GPU memory issues. Through a coordinate descent strategy and a centrality-enforcing activation function, DARC produces unbiased, diffeomorphic atlases with high anatomical fidelity. Beyond atlas construction, we demonstrate two key applications: (1) One-shot segmentation, where labels annotated only on the atlas are propagated to subjects via inverse deformations, outperforming state-of-the-art few-shot methods; and (2) shape synthesis, where new anatomical variants are generated by warping the atlas mesh using synthesized diffeomorphic deformation fields. Overall, DARC offers a flexible, generalizable, and resource-efficient framework for atlas construction and applications.

Problem

• Develops efficient model-driven groupwise atlas construction method
• Addresses scalability and optimization in 3D medical image registration
• Enables one-shot segmentation and shape synthesis applications

Innovation

• Model-driven groupwise registration framework DARC
• Coordinate descent strategy for unbiased atlases
• Supports diverse metrics without GPU memory issues