Deep Learning Automatic Fetal Structures Segmentation in MRI Scans with Few Annotated Datasets

Abstract

We present a new method for end-to-end automatic volumetric segmentation of fetal structures in MRI scans with deep learning networks trained with very few annotated scans. It consists of three main stages: 1) two-step automatic structure segmentation with custom 3D U-Nets; 2) segmentation error estimation, and; 3) segmentation error correction. The automatic structure segmentation stage first computes a region of interest (ROI) on a downscaled scan and then computes a final segmentation on the cropped ROI. The segmentation error estimation stage uses prediction-time augmentations of the input scan to compute multiple segmentations and estimate the segmentation uncertainty for individual slices and for the entire scan. The segmentation error correction stage then uses these estimations to locate the most error-prone slices and to correct the segmentations in those slices based on validated adjacent slices. Experimental results of our methods on fetal body (63 cases, 9 for training, 55 for testing) and fetal brain MRI scans (35 cases, 6 for training, 29 for testing) yield a mean Dice coefficient of 0.96 for both, and a mean Average Symmetric Surface Distance of 0.74 mm and 0.19 mm, respectively, below the observer delineation variability.