Evolving convolutional neural network parameters through the genetic algorithm for the breast cancer classification...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Breast cancer is a leading cause of cancer mortality in women worldwide.
• Early diagnosis and effective treatment are crucial for controlling breast cancer.
• Computer-aided diagnosis (CAD) systems enhance diagnostic accuracy for clinical specialists.

Purpose of the Study:

• To optimize the weights of a convolutional neural network (CNN) for breast cancer classification using the genetic algorithm (GA).
• To compare the performance of GA-optimized CNNs against traditional optimizers like mini-batch gradient descent and Adam.
• To evaluate the effectiveness of deep learning techniques in improving breast cancer diagnosis accuracy.

Main Methods:

• Designing a CNN model for breast cancer image classification.
• Training the CNN model using three different optimizers: mini-batch gradient descent, Adam, and GA.
• Evaluating model performance on the BreakHis dataset through rigorous experiments.

Main Results:

• The CNN model trained with the GA achieved a classification accuracy of 85%.
• The GA optimizer performed comparably to the Adam optimizer in terms of classification accuracy.
• The study demonstrates the potential of GA for optimizing CNN parameters in medical image analysis.

Conclusions:

• The genetic algorithm (GA) is a viable and effective optimizer for training convolutional neural networks (CNNs) in breast cancer diagnosis.
• GA-optimized CNNs can achieve high classification accuracy, comparable to advanced optimizers like Adam.
• This approach offers a promising alternative for improving the accuracy and efficiency of computer-aided diagnosis (CAD) systems in oncology.