An evaluation of U-Net in Renal Structure Segmentation

TL;DR

This paper evaluates various U-Net models for renal structure segmentation in CTA images, highlighting their performance in the KiPA 2022 challenge to advance kidney imaging analysis.

Contribution

The study systematically assesses different U-Net variants on a new multi-structure renal dataset, identifying the most effective models for segmentation tasks.

Findings

U-Net variants achieved high segmentation accuracy.

Best models significantly improved renal structure delineation.

Evaluation results inform future medical image segmentation approaches.

Abstract

Renal structure segmentation from computed tomography angiography~(CTA) is essential for many computer-assisted renal cancer treatment applications. Kidney PArsing~(KiPA 2022) Challenge aims to build a fine-grained multi-structure dataset and improve the segmentation of multiple renal structures. Recently, U-Net has dominated the medical image segmentation. In the KiPA challenge, we evaluated several U-Net variants and selected the best models for the final submission.