Optimized Deep Learning Model for Early Detection of Breast Cancer in Mammograms Using Vision Transformer
International Journal of Drug Delivery Technology · 30.04.2026 · 16 (23s)
DOI: 10.25258/ijddt.16.23s.2 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 30.04.2026. Verfasst von Daphne Sherine. H und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Breast cancer (BC) is a major global health issue and one of the leading causes of cancer-related mortality for women. A timely and precise diagnosis using mammography analysis is essential for increasing survival rates. With the use of Vision Transformers' (ViTs) potent representational capabilities, this research presents an improved deep learning (DL) framework for the early identification and classification of BC in mammograms. The proposed model, Artificial Jellyfish Search-Driven Attention-Enriched Convolutional tuned Pyramid Vision Transformer (AJFS-Att-CPVT), is designed to detect and classify BC by effectively capturing both global and local image features to distinguish between benign and malignant lesions. Mammographic data are collected from multiple publicly available datasets to ensure diversity in breast tissue characteristics and lesion types. The preprocessing pipeline included image normalization and Gaussian blurring to standardize image quality and improve feature visibility. To strengthen model generalization and mitigate overfitting, data augmentation strategies such as random rotation, horizontal flipping, and contrast variation were employed during training. For baseline feature extraction, ResNet50 was applied to the preprocessed images, providing a benchmark for evaluating the performance gains achieved by the proposed AJFS-Att-CPVT model. Model evaluation using Python with key performance metrics, such as a recall of 94.8%, an F1-score of 93.8%, an accuracy of 95.8%, a specificity of 94.20%, an AUC of 93.35%, and a precision of 95.2%, demonstrated higher diagnostic capabilities compared to traditional convolutional approaches. The findings demonstrate how Vision Transformer-based designs have the potential to greatly improve the accuracy and dependability of early BC diagnosis in mammography imaging when refined with bio-inspired algorithms and sophisticated preprocessing.
Zitieren
APA
H, D. S., & Revathy, G. (2026). Optimized Deep Learning Model for Early Detection of Breast Cancer in Mammograms Using Vision Transformer. International Journal of Drug Delivery Technology, 16(23s). https://doi.org/10.25258/ijddt.16.23s.2
Vancouver
H DS, Revathy G. Optimized Deep Learning Model for Early Detection of Breast Cancer in Mammograms Using Vision Transformer. International Journal of Drug Delivery Technology. 2026;16(23s). doi:10.25258/ijddt.16.23s.2
BibTeX
@article{H2026,
author = {H, Daphne Sherine. and Revathy, G.},
title = {Optimized Deep Learning Model for Early Detection of Breast Cancer in Mammograms Using Vision Transformer},
journal = {International Journal of Drug Delivery Technology},
year = {2026},
volume = {16},
number = {23s},
doi = {10.25258/ijddt.16.23s.2}
}
EUVIMED