KiPA22 Report: U-Net with Contour Regularization for Renal Structures Segmentation

TL;DR

This paper presents a modified nnU-Net framework with contour regularization loss to improve 3D renal structure segmentation accuracy, especially reducing tumor label outliers, demonstrating advancements in medical image segmentation techniques.

Contribution

The paper introduces a novel contour regularization loss integrated with nnU-Net for better renal structure segmentation, focusing on tumor label accuracy.

Findings

Improved segmentation accuracy for renal structures.

Reduced outlier predictions for tumor labels.

Enhanced robustness of the segmentation model.

Abstract

Three-dimensional (3D) integrated renal structures (IRS) segmentation is important in clinical practice. With the advancement of deep learning techniques, many powerful frameworks focusing on medical image segmentation are proposed. In this challenge, we utilized the nnU-Net framework, which is the state-of-the-art method for medical image segmentation. To reduce the outlier prediction for the tumor label, we combine contour regularization (CR) loss of the tumor label with Dice loss and cross-entropy loss to improve this phenomenon.