Transformers and Hybrid AI Models for Accurate Breast Cancer Segmentation

DOI

10.2991/978-94-6463-704-5_17How to use a DOI?

Keywords

Dice Similarity Coefficient (DSC); Transformer-CNN Integration; Deep Learning (DL)

Abstract

Precise segmentation of breast cancer in imaging is critical for diagnosis, treatment planning, and outcome prediction in medical images. The inherent complexity of tumor shape, heterogeneous boundary and multi-modal imaging data have made traditional segmentation techniques based on classical image processing as well as early deep learning models (like U-Net) inadequate. In this paper we present a new segmentation framework that combines transformer-based architectures with CNNs taking advantage of their complementary strength i.e. local feature extraction by CNNs and global context modeling by transformers. We also assess a proposed hybrid breast segmentation model, the Trans-CNN Breast Segmentation Network (T-CBSN), on diverse datasets containing mammograms and magnetic resonance imaging (MRI) scans. Robustness under different imaging conditions was achieved through advanced preprocessing (e.g., denoising and augmentation). The model exceeded performance with a DSC of 96% rather than state-of-the-art U-Net (92%) and Swin U-Net (94%). Other metrics, such as sensitivity (97%) and specificity (95%), supported its reliability in detecting true tumor regions while limiting false positives.

Copyright

© 2025 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  - Kranti K. Dewangan
AU  - Satya Prakash Sahu
AU  - Rekh Ram Janghel
PY  - 2025
DA  - 2025/04/30
TI  - Transformers and Hybrid AI Models for Accurate Breast Cancer Segmentation
BT  - Proceedings of the International Conference on Smart Health and Intelligent Technologies (ICSHit-2024)
PB  - Atlantis Press
SP  - 223
EP  - 234
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-704-5_17
DO  - 10.2991/978-94-6463-704-5_17
ID  - Dewangan2025
ER  -