Enhancing NnU-Net for Improved Medical Image Segmentation: A Comparative Study with TotalSegmentator

Jingyi Wu

2024

Abstract

In this paper, an optimization method is proposed that relies on the no-new-Net (nnU Net) architecture to improve the performance of medical image segmentation tasks. Medical image segmentation is an important component of disease diagnosis, treatment planning, and surgical assistance. Since its launch in 2018, nnU Net has become a fundamental tool in this field by adapting its architecture, preprocessing, and training strategies. However, current models still have shortcomings in handling data imbalance and multimodal images. For this purpose, the paper optimized the loss function and data augmentation strategy of nnU Net. By increasing the Dice loss weight, the model can more effectively handle small structures and imbalanced data, improving segmentation accuracy. Furthermore, by incorporating higher rotation probability, noise enhancement, and low-resolution simulation into the improved data augmentation technique, the model's robustness and capacity for generalization are greatly increased. The experimental results demonstrate that the upgraded nnU Net performs much better than TotalSegmentor in terms of segmentation accuracy and complicated boundary handling, especially when compared to metrics like Dice Score, IoU, and Hausdorff Distance.

Paper Citation

in Harvard Style

Wu J. (2024). Enhancing NnU-Net for Improved Medical Image Segmentation: A Comparative Study with TotalSegmentator. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 246-251. DOI: 10.5220/0013515100004619

in Bibtex Style

@conference{daml24,
author={Jingyi Wu},
title={Enhancing NnU-Net for Improved Medical Image Segmentation: A Comparative Study with TotalSegmentator},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={246-251},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013515100004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Enhancing NnU-Net for Improved Medical Image Segmentation: A Comparative Study with TotalSegmentator
SN - 978-989-758-754-2
AU - Wu J.
PY - 2024
SP - 246
EP - 251
DO - 10.5220/0013515100004619
PB - SciTePress