Learn to Segment Organs with a Few Bounding Boxes

TL;DR

This paper introduces a U-Net based method that performs organ segmentation in 3D medical images using only a few bounding box annotations, reducing the need for extensive pixel-level labeled datasets.

Contribution

It presents a novel few-shot segmentation approach that leverages bounding boxes and prototype learning to segment unseen organs with minimal annotations.

Findings

Achieved median segmentation score of 54.64% on unseen organs.

Reduces annotation effort by using bounding boxes instead of full labels.

Enables fast 3D organ segmentation with limited supervision.

Abstract

Semantic segmentation is an import task in the medical field to identify the exact extent and orientation of significant structures like organs and pathology. Deep neural networks can perform this task well by leveraging the information from a large well-labeled data-set. This paper aims to present a method that mitigates the necessity of an extensive well-labeled data-set. This method also addresses semi-supervision by enabling segmentation based on bounding box annotations, avoiding the need for full pixel-level annotations. The network presented consists of a single U-Net based unbranched architecture that generates a few-shot segmentation for an unseen human organ using just 4 example annotations of that specific organ. The network is trained by alternately minimizing the nearest neighbor loss for prototype learning and a weighted cross-entropy loss for segmentation learning to perform a fast 3D segmentation with a median score of 54.64%.