ITDS | SPMS

A NOVEL APPROACH FOR EARLY DETECTION OF BREAST CANCER USING ONE-CLASS CLASSIFICATION AND DEEP LEARNING TECHNIQUES: ADDRESSING CLASS IMBALANCE

Project Description
Supervisor

Breast cancer continues to be a leading cause of mortality, for women. Emphasizes the need for improved and timely detection methods. This study introduces an approach to diagnosing breast cancer by integrating one-class classification (OCC) methods with advanced deep learning technologies, like MobileNet, DenseNet169, and Xception models. Traditional methods for detecting breast cancer frequently rely on multi-class classifiers, which suffer from data imbalance, particularly in medical imaging, where normal instances outnumber suspicious ones. In contrast, OCC excels in detecting anomalies in normal data, making it ideal for early-stage breast cancer detection. The study used mammography, ultrasound, and MRI images, with models trained mostly on normal breast tissue. An early stopping mechanism with patience of 5 was employed to prevent overfitting, ensuring optimal performance during model training. The study's major goal was to assess the accuracy, sensitivity, and specificity of different OCC models, including as MobileNet, DenseNet169, and Xception, in recognizing benign and malignant breast tumors. The models were assessed based on precision, recall, and F1-score metrics. The MobileNet model achieved an accuracy of 89%, while DenseNet169 and Xception yielded comparable results, with 88% and 87% accuracy, respectively. These models exhibited strong detection capabilities for both benign and malignant cases, surpassing numerous other methods documented in the literature. Additionally, saliency maps were utilized to enhance the interpretability of the models by providing visual insights into the areas of the image that significantly impacted the models' decision-making processes. The study demonstrates that One-Class Classification (OCC) combined with deep learning is effective for breast cancer detection, but suggests future research on model generalization and dataset diversity.

Machine Learning: Machine Learning (ML) research in Computer Science and Information Technology focuses on the development of algorithms and models that enable computers to learn from data and improve their performance over time without being explicitly programmed. It is a subset of Artificial Intelligence that uses statistical techniques to give machines the ability to learn patterns, make decisions, and predict outcomes based on data. Supervised learning, a key area of ML research, involves training models on labeled data, where the input-output relationships are predefined. This method is widely used for tasks such as classification (e.g., spam detection) and regression (e.g., predicting house prices). Unsupervised learning, on the other hand, involves finding hidden patterns in data without predefined labels, with clustering and association being typical applications in areas such as customer segmentation and anomaly detection. Reinforcement learning is another area of ML that focuses on teaching agents to make decisions by interacting with their environment and receiving feedback in the form of rewards or penalties. It is often applied in robotics, game playing, and autonomous systems, where continuous learning and adaptation are required.