3D Cross-Pseudo Supervision (3D-CPS): A semi-supervised nnU-Net architecture for abdominal organ segmentation

TL;DR

This paper introduces 3D-CPS, a semi-supervised nnU-Net architecture that effectively leverages unlabeled medical images for abdominal organ segmentation, achieving high accuracy with reduced annotation effort.

Contribution

The paper presents a novel 3D Cross-Pseudo Supervision method integrated into nnU-Net, with a new preprocessing and loss weighting strategy to improve semi-supervised segmentation performance.

Findings

Achieved DSC of 0.881 on MICCAI FLARE2022 validation set

Achieved NSD of 0.913 on validation set

Demonstrated effectiveness of semi-supervised approach in medical image segmentation

Abstract

Large curated datasets are necessary, but annotating medical images is a time-consuming, laborious, and expensive process. Therefore, recent supervised methods are focusing on utilizing a large amount of unlabeled data. However, to do so, is a challenging task. To address this problem, we propose a new 3D Cross-Pseudo Supervision (3D-CPS) method, a semi-supervised network architecture based on nnU-Net with the Cross-Pseudo Supervision method. We design a new nnU-Net based preprocessing. In addition, we set the semi-supervised loss weights to expand linearity with each epoch to prevent the model from low-quality pseudo-labels in the early training process. Our proposed method achieves an average dice similarity coefficient (DSC) of 0.881 and an average normalized surface distance (NSD) of 0.913 on the MICCAI FLARE2022 validation set (20 cases).