Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning

The increasing demand for accurate and efficient medical image segmentation has driven the development of advanced Artificial Intelligence (AI) models to assist in diagnosing and planning treatments. In this paper, we propose a novel approach that combines U-Net architecture with attention mechanism...

Full description

Saved in:
Bibliographic Details
Main Authors: Neha Tirpude, Tausif Diwan, Meera Dhabu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10960382/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849699868658892800
author Neha Tirpude
Tausif Diwan
Meera Dhabu
author_facet Neha Tirpude
Tausif Diwan
Meera Dhabu
author_sort Neha Tirpude
collection DOAJ
description The increasing demand for accurate and efficient medical image segmentation has driven the development of advanced Artificial Intelligence (AI) models to assist in diagnosing and planning treatments. In this paper, we propose a novel approach that combines U-Net architecture with attention mechanisms and Model-Agnostic Meta-Learning (MAML) to enhance segmentation performance, particularly in scenarios with limited annotated data. By applying this method to the AMOS dataset for multi-organ segmentation in CT and MRI scans, our model significantly improves generalization across different patients and imaging modalities. The use of attention mechanisms refines feature extraction, while class weighting effectively addresses class imbalance. The proposed model achieves a state-of-the-art Dice score of 98.09%, pixel accuracy of 98.39%, precision of 98.58% and recall of 98.27% outperforming existing methods and demonstrating the potential for rapid adaptation in clinical applications. This combination of meta-learning and deep learning offers a robust solution for addressing the challenges of medical image segmentation.
format Article
id doaj-art-98d59bcbd8bb43c2a6e6379a5f155c7e
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-98d59bcbd8bb43c2a6e6379a5f155c7e2025-08-20T03:18:27ZengIEEEIEEE Access2169-35362025-01-0113648226483110.1109/ACCESS.2025.355944610960382Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-LearningNeha Tirpude0https://orcid.org/0000-0002-8830-2630Tausif Diwan1Meera Dhabu2Department of Computer Science and Engineering, Indian Institute of Information Technology, Nagpur, IndiaDepartment of Computer Science and Engineering, Indian Institute of Information Technology, Nagpur, IndiaDepartment of Computer Science and Engineering, Visvesvaraya National Institute of Technology, Nagpur, IndiaThe increasing demand for accurate and efficient medical image segmentation has driven the development of advanced Artificial Intelligence (AI) models to assist in diagnosing and planning treatments. In this paper, we propose a novel approach that combines U-Net architecture with attention mechanisms and Model-Agnostic Meta-Learning (MAML) to enhance segmentation performance, particularly in scenarios with limited annotated data. By applying this method to the AMOS dataset for multi-organ segmentation in CT and MRI scans, our model significantly improves generalization across different patients and imaging modalities. The use of attention mechanisms refines feature extraction, while class weighting effectively addresses class imbalance. The proposed model achieves a state-of-the-art Dice score of 98.09%, pixel accuracy of 98.39%, precision of 98.58% and recall of 98.27% outperforming existing methods and demonstrating the potential for rapid adaptation in clinical applications. This combination of meta-learning and deep learning offers a robust solution for addressing the challenges of medical image segmentation.https://ieeexplore.ieee.org/document/10960382/AMOS datasetattention mechanismsmedical image segmentationmeta-learningU-Net architecture
spellingShingle Neha Tirpude
Tausif Diwan
Meera Dhabu
Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
IEEE Access
AMOS dataset
attention mechanisms
medical image segmentation
meta-learning
U-Net architecture
title Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
title_full Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
title_fullStr Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
title_full_unstemmed Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
title_short Meta-Transformers: A Hybrid Approach for Medical Image Segmentation With U-Net and Meta-Learning
title_sort meta transformers a hybrid approach for medical image segmentation with u net and meta learning
topic AMOS dataset
attention mechanisms
medical image segmentation
meta-learning
U-Net architecture
url https://ieeexplore.ieee.org/document/10960382/
work_keys_str_mv AT nehatirpude metatransformersahybridapproachformedicalimagesegmentationwithunetandmetalearning
AT tausifdiwan metatransformersahybridapproachformedicalimagesegmentationwithunetandmetalearning
AT meeradhabu metatransformersahybridapproachformedicalimagesegmentationwithunetandmetalearning