Multi-Scale Attention Transformer Network for Robust Brain Tumor Segmentation across MRI Modalities

DOI

10.2991/978-94-6239-674-6_20How to use a DOI?

Keywords

Brain tumor segmentation; MRI; deep learning; transformer network; multi-scale attention; medical image analysis; MSAT-Net; and multimodal fusion

Abstract

The objective of this paper is to introduce Multi-Scale Attention Transformer Network (MSAT-Net) for the purpose of providing accurate and reliable brain tumor segmentation over various MRI datasets. This makes it easier to characterize tumor subregions such as edema, necrotic core and enhancing tissues. The transformer encoder combines multimodal MRI inputs T1, T1c, T2, and FLAIR among others into a universal consistent representation. Tests conducted on benchmark datasets indicates that MSAT-Net performs better than baseline CNNs and transformer hybrids with regards to Dice score, sensitivity, boundary accuracy.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Prashant Dixit
AU  - Kumud Dixit
AU  - Ankit Upadhyay
AU  - Abhishek Varshney
AU  - Sujeet Kumar
AU  - Methly Johri
PY  - 2026
DA  - 2026/05/28
TI  - Multi-Scale Attention Transformer Network for Robust Brain Tumor Segmentation across MRI Modalities
BT  - Proceedings of the International Conference on Sustainable Computing and Artificial Intelligence (ICSCAI 2025)
PB  - Atlantis Press
SP  - 229
EP  - 242
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-674-6_20
DO  - 10.2991/978-94-6239-674-6_20
ID  - Dixit2026
ER  -