Enhancing Single-Slice Segmentation with 3D-to-2D Unpaired Scan Distillation

TL;DR

This paper introduces a 3D-to-2D distillation framework that leverages unpaired 3D CT scans to improve 2D single-slice segmentation accuracy, especially in low-data scenarios, reducing manual annotation efforts.

Contribution

The novel framework uses unpaired 3D models to guide 2D segmentation, enhancing performance without requiring paired data or extensive annotations.

Findings

Improved multi-organ segmentation accuracy on BLSA dataset.

Significant gains in low-data regimes with only 200 training subjects.

Outperforms models trained on all available data in certain conditions.

Abstract

2D single-slice abdominal computed tomography (CT) enables the assessment of body habitus and organ health with low radiation exposure. However, single-slice data necessitates the use of 2D networks for segmentation, but these networks often struggle to capture contextual information effectively. Consequently, even when trained on identical datasets, 3D networks typically achieve superior segmentation results. In this work, we propose a novel 3D-to-2D distillation framework, leveraging pre-trained 3D models to enhance 2D single-slice segmentation. Specifically, we extract the prediction distribution centroid from the 3D representations, to guide the 2D student by learning intra- and inter-class correlation. Unlike traditional knowledge distillation methods that require the same data input, our approach employs unpaired 3D CT scans with any contrast to guide the 2D student model. Experiments conducted on 707 subjects from the single-slice Baltimore Longitudinal Study of Aging (BLSA) dataset demonstrate that state-of-the-art 2D multi-organ segmentation methods can benefit from the 3D teacher model, achieving enhanced performance in single-slice multi-organ segmentation. Notably, our approach demonstrates considerable efficacy in low-data regimes, outperforming the model trained with all available training subjects even when utilizing only 200 training subjects. Thus, this work underscores the potential to alleviate manual annotation burdens.