Conditional Fetal Brain Atlas Learning for Automatic Tissue Segmentation
Authors
Abstract
Magnetic Resonance Imaging (MRI) of the fetal brain has become a key tool for studying brain development in vivo. Yet, its assessment remains challenging due to variability in brain maturation, imaging protocols, and uncertain estimates of Gestational Age (GA). To overcome these, brain atlases provide a standardized reference framework that facilitates objective evaluation and comparison across subjects by aligning the atlas and subjects in a common coordinate system. In this work, we introduce a novel deep-learning framework for generating continuous, age-specific fetal brain atlases for real-time fetal brain tissue segmentation. The framework combines a direct registration model with a conditional discriminator. Trained on a curated dataset of 219 neurotypical fetal MRIs spanning from 21 to 37 weeks of gestation. The method achieves high registration accuracy, captures dynamic anatomical changes with sharp structural detail, and robust segmentation performance with an average Dice Similarity Coefficient (DSC) of 86.3% across six brain tissues. Furthermore, volumetric analysis of the generated atlases reveals detailed neurotypical growth trajectories, providing valuable insights into the maturation of the fetal brain. This approach enables individualized developmental assessment with minimal pre-processing and real-time performance, supporting both research and clinical applications. The model code is available at https://github.com/cirmuw/fetal-brain-atlas