Related Experiment Video
Updated: May 13, 2026
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
BC-SwinNet: Swin transformer and CNN with multi-objective optimization for multi-class breast cancer detection using
Mudassir Khan1, Meteb Altaf2, Nikhat Raza Khan3
1Department of Informatics and Computer Systems, College of Computer Science, King Khalid University, Abha, Saudi Arabia.
Scientific Reports
|May 11, 2026
Summary
A new deep learning model, BC-SwinNet, accurately classifies breast cancer subtypes from histopathology images. This automated system shows high performance, offering potential improvements for clinical diagnostic support.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Breast cancer is a leading cause of death in women, with rising incidence projected by 2030.
- Manual histopathological analysis is subjective and time-consuming, necessitating automated diagnostic solutions.
- Accurate classification of breast cancer subtypes is crucial for effective treatment planning.
Purpose of the Study:
- To develop an automated deep learning framework for classifying multiple breast cancer subtypes using histopathological images.
- To address the limitations of manual diagnosis, including subjectivity and inter-observer variability.
- To enhance the accuracy and efficiency of breast cancer diagnosis in clinical settings.
Main Methods:
- Introduction of BC-SwinNet, a hybrid deep learning framework combining Conditional Swin Transformer (ConSwinTras), Multi-Objective Elk Herd Optimization (MEHO), and Layered Attention-based Convolutional Neural Network (CA-CNN).
- ConSwinTras for hierarchical and contextual feature extraction.
- MEHO for feature selection and dimensionality reduction.
- CA-CNN with layer-wise attention for focused analysis and classification.
Main Results:
- BC-SwinNet achieved high classification accuracies of 99.91% on the BreakHis dataset and 99.854% on the BACH dataset.
- The proposed framework demonstrated superior performance compared to existing methods.
- The model maintained computational efficiency while achieving high diagnostic accuracy.
Conclusions:
- BC-SwinNet provides a robust and efficient automated approach for breast cancer subtype classification.
- The framework has the potential to significantly enhance clinical decision-making and diagnostic support systems.
- Automated analysis of histopathological images using BC-SwinNet can improve the accuracy and speed of breast cancer diagnosis.
