Breast Cancer Classification Depends on the Dynamic Dipper Throated Optimization Algorithm.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Breast cancer is a leading cause of mortality in women, necessitating early detection.
• Manual diagnosis of breast cancer is time-consuming, highlighting the need for automated methods.
• Early diagnosis and treatment are crucial for reducing breast cancer death rates.

Purpose of the Study:

• To develop a robust automated framework for breast cancer classification from ultrasound images.
• To integrate metaheuristic optimization, deep learning, and feature selection for enhanced diagnostic accuracy.
• To improve upon existing methods for early breast cancer identification.

Main Methods:

• A five-stage framework including data augmentation for Convolutional Neural Network (CNN) models.
• Transfer learning using GoogleNet for feature extraction.
• A novel hybrid optimization algorithm (dipper throated and particle swarm optimization) for feature selection.
• Classification using a CNN optimized with the proposed hybrid algorithm.

Main Results:

• The proposed framework achieved a classification accuracy of 98.1% on a public breast cancer dataset.
• The novel feature selection method and optimized CNN demonstrated superior performance.
• Statistical tests confirmed the stability and effectiveness compared to state-of-the-art approaches.

Conclusions:

• The proposed integrated framework offers a highly accurate and efficient method for automated breast cancer classification.
• This approach holds significant potential for improving early breast cancer diagnosis from ultrasound images.
• The optimized deep learning model and feature selection technique represent a significant advancement in medical image analysis for oncology.