# PAN: Projective Adversarial Network for Medical Image Segmentation

## TL;DR

This paper introduces PAN, a novel adversarial network that efficiently captures 3D semantics in medical image segmentation by using 2D projections and an attention module, achieving state-of-the-art results in pancreas segmentation.

## Contribution

The paper proposes a projective adversarial network with an attention module that effectively incorporates 3D information in medical image segmentation without increasing segmentor complexity.

## Key findings

- Achieved state-of-the-art pancreas segmentation performance.
- Efficiently captures 3D semantics using 2D projections.
- Enhances segmentation accuracy with attention-guided global information integration.

## Abstract

Adversarial learning has been proven to be effective for capturing long-range and high-level label consistencies in semantic segmentation. Unique to medical imaging, capturing 3D semantics in an effective yet computationally efficient way remains an open problem. In this study, we address this computational burden by proposing a novel projective adversarial network, called PAN, which incorporates high-level 3D information through 2D projections. Furthermore, we introduce an attention module into our framework that helps for a selective integration of global information directly from our segmentor to our adversarial network. For the clinical application we chose pancreas segmentation from CT scans. Our proposed framework achieved state-of-the-art performance without adding to the complexity of the segmentor.