KiPA22 Report: U-Net with Contour Regularization for Renal Structures Segmentation
This work addresses segmentation accuracy for renal structures in medical imaging, but it is incremental as it builds on the established nnU-Net with a minor modification.
The authors tackled the problem of 3D integrated renal structures segmentation, particularly reducing outlier predictions for tumor labels, by combining contour regularization loss with Dice and cross-entropy losses in the nnU-Net framework, achieving unspecified improvements.
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.