ABCanDroid: A Cloud Integrated Android App for Noninvasive Early Breast Cancer Detection Using Transfer Learning

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Improper diagnosis and treatment of breast cancer lead to significant mortality.
• Deep learning shows promise in breast cancer detection but requires further optimization.
• Transfer learning offers a pathway to improve the efficiency and accuracy of these methods.

Purpose of the Study:

• To enhance the accuracy and efficiency of early breast cancer detection.
• To leverage transfer learning with Convolutional Neural Networks (CNNs) for improved diagnostic capabilities.
• To develop a robust framework for early breast cancer identification.

Main Methods:

• Utilizing a pre-trained ResNet101 model, a deep learning architecture, for transfer learning.
• Employing the ImageNet dataset to initialize model weights, avoiding training from scratch.
• Integrating Convolutional Neural Network (CNN) principles with transfer learning techniques.

Main Results:

• The proposed ResNet101-based transfer learning framework achieved a high classification accuracy of 99.58%.
• Extensive experiments and hyperparameter tuning were conducted to optimize performance.
• The model demonstrated significant potential for accurate breast cancer classification.

Conclusions:

• The developed framework offers a highly accurate and efficient tool for early breast cancer detection.
• Transfer learning significantly boosts the performance of deep learning models in this domain.
• This approach can serve as a valuable aid for clinicians and improve patient outcomes.