Placenta Segmentation in Ultrasound Imaging: Addressing Sources of Uncertainty and Limited Field-of-View

TL;DR

This paper presents a multi-task neural network approach for placenta segmentation in ultrasound images, addressing variability, limited annotations, and field-of-view constraints to achieve human-level accuracy and whole-structure assessment.

Contribution

The work introduces a multi-task learning method combining placenta classification and segmentation, enabling improved accuracy and whole-placenta segmentation from multi-view ultrasound data.

Findings

Achieved Dice scores of 0.86 and 0.83 for anterior and posterior placentas.

Model performance comparable to human intra- and inter-observer variability.

Enabled whole placenta segmentation using multi-view ultrasound pipeline.

Abstract

Automatic segmentation of the placenta in fetal ultrasound (US) is challenging due to the (i) high diversity of placenta appearance, (ii) the restricted quality in US resulting in highly variable reference annotations, and (iii) the limited field-of-view of US prohibiting whole placenta assessment at late gestation. In this work, we address these three challenges with a multi-task learning approach that combines the classification of placental location (e.g., anterior, posterior) and semantic placenta segmentation in a single convolutional neural network. Through the classification task the model can learn from larger and more diverse datasets while improving the accuracy of the segmentation task in particular in limited training set conditions. With this approach we investigate the variability in annotations from multiple raters and show that our automatic segmentations (Dice of 0.86 for anterior and 0.83 for posterior placentas) achieve human-level performance as compared to intra- and inter-observer variability. Lastly, our approach can deliver whole placenta segmentation using a multi-view US acquisition pipeline consisting of three stages: multi-probe image acquisition, image fusion and image segmentation. This results in high quality segmentation of larger structures such as the placenta in US with reduced image artifacts which are beyond the field-of-view of single probes.