Enhancing Breast Cancer Detection in Ultrasound Images-An Innovative Approach Using Progressive Fine‐Tuning of Vision Transformer Models
Abstract
Breast cancer is ranked as the second most common cancer among women globally, highlighting the critical need for precise and
early detection methods. Our research introduces a novel approach for classifying benign and malignant breast ultrasound images.
We leverage advanced deep learning methodologies, mainly focusing on the vision transformer (ViT) model. Our method
distinctively features progressive 4ne-tuning, a tailored process that incrementally adapts the model to the nuances of breast tissue
classi4cation. Ultrasound imaging was chosen for its distinct bene4ts in medical diagnostics. )is modality is noninvasive and
cost-e7ective and demonstrates enhanced speci4city, especially in dense breast tissues where traditional methods may struggle.
Such characteristics make it an ideal choice for the sensitive task of breast cancer detection. Our extensive experiments utilized the
breast ultrasound images dataset, comprising 780 images of both benign and malignant breast tissues. )e dataset underwent
a comprehensive analysis using several pretrained deep learning models, including VGG16, VGG19, DenseNet121, Inception,
ResNet152V2, DenseNet169, DenseNet201, and the ViT. )e results presented were achieved without employing data augmentation
techniques. )e ViTmodel demonstrated robust accuracy and generalization capabilities with the original dataset size,
which consisted of 637 images. Each model’s performance was meticulously evaluated through a robust 10-fold cross-validation
technique, ensuring a thorough and unbiased comparison. Our 4ndings are signi4cant, demonstrating that the progressive 4netuning
substantially enhances the ViT model’s capability. )is resulted in a remarkable accuracy of 94.49% and an AUC score of
0.921, signi4cantly higher than models without 4ne-tuning. )ese results aErm the eEcacy of the ViT model and highlight the
transformative potential of integrating progressive 4ne-tuning with transformer models in medical image classi4cation tasks. )e
study solidi4es the role of such advanced methodologies in improving early breast cancer detection and diagnosis, especially when
coupled with the unique advantages of ultrasound imaging.