This project aims to classify breast cancer using Convolutional Neural Networks (CNN) and ResNet, comparing their performance based on accuracy, recall, F1-score, precision, and Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve.
Breast cancer is one of the most prevalent forms of cancer among women worldwide. Early detection is crucial for effective treatment and improved survival rates.
Machine learning techniques, particularly deep learning models, have shown promising results in automating the detection and classification of breast cancer from medical images.
In this project, I employ CNN and ResNet architectures to classify breast cancer from mammogram images. The models are evaluated on multiple metrics to determine their effectiveness in accurately diagnosing breast cancer.
I utilize the Breast Cancer Histopathological Database (BreakHis) for training and evaluation.
This dataset consists of high-resolution histopathological images of breast tissue samples labeled with cancerous and non-cancerous regions.
I design a custom Convolutional Neural Network (CNN) architecture tailored for breast cancer classification.
The CNN comprises multiple convolutional, activation, and pooling layers, followed by fully connected layers for classification.
I leverage the Residual Neural Network (ResNet) architecture, known for its deep structure and skip connections, which facilitate training of very deep networks.
A pre-trained ResNet model is fine-tuned on the breast cancer dataset for optimized classification performance.
The performance of both CNN and ResNet models is evaluated using the following metrics:
- Accuracy – Overall proportion of correctly classified samples
- Recall – Proportion of actual positive cases correctly identified
- Precision – Proportion of predicted positive cases that are correct
- F1-score – Harmonic mean of precision and recall
- AUC (ROC) – Measure of the model's ability to distinguish between classes across various thresholds
The comparative performance of CNN and ResNet models is analyzed based on the above metrics.
ROC curves are plotted to visualize the trade-off between true positive rate and false positive rate, showcasing each model’s ability to discriminate between cancerous and non-cancerous samples.
I acknowledge the Breast Cancer Histopathological Database (BreakHis) for providing the dataset used in this project.