ORIGINAL RESEARCH article
Front. Oncol.
Sec. Cancer Imaging and Image-directed Interventions
Volume 15 - 2025 | doi: 10.3389/fonc.2025.1467672
This article is part of the Research TopicDeep Learning for Medical Imaging ApplicationsView all 15 articles
UnetTransCNN: Integrating Transformers with Convolutional Neural Networks for Enhanced Medical Image Segmentation
Provisionally accepted
- Inner Mongolia University, Hohhot, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
This paper proposes UnetTransCNN, a novel model that integrates Transformers with Convolutional Neural Networks (CNNs) to enhance medical image segmentation. Our approach leverages Transformers for global context capture and CNNs for local feature extraction to improve segmentation accuracy and efficiency. UnetTransCNN introduces a parallel architecture that extracts both local and global features, coupled with adaptive global-local coupling units for optimal feature fusion. Experiments on benchmark datasets, including the Multi Atlas Labeling Beyond The Cranial Vault (BTCV) and the Medical Segmentation Decathlon (MSD), demonstrate that UnetTransCNN achieves state-of-the-art performance, particularly excelling in both large and small organ segmentation tasks. Our model shows superior metrics and robustness across various hyperparameters, positioning it as an advancement in medical image segmentation.Specifically, UnetTransCNN remarkably surpasses the second-best baselines by considerable margins of 6.382% and 6.772% in the Dice score for the gallbladder and adrenal glands, respectively.
Keywords: Fully Convolutional Neural Networks, transformer, Medical image segmentation, 3D image, Feature fusion
Received: 23 Jul 2024; Accepted: 15 Apr 2025.
Copyright: © 2025 Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Fan Li, Inner Mongolia University, Hohhot, China
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.