TrueScan – Breast Cancer Detection with Machine Learning

Breast cancer is a critical health concern globally, ranking among the most significant causes of mortality in women. Timely and accurate detection substantially improves treatment effectiveness and survival rates. However, conventional diagnostic methods, which depend primarily on manual examination of histopathology slides, are labor-intensive, timeconsuming, and susceptible to subjective interpretation and human error. This study introduces TrueScan, an AI-powered framework designed to facilitate breast cancer detection through machine learning (ML) and deep learning (DL) techniques. The system employs a Convolutional Neural Network (CNN) as the primary classification model for histopathology images and incorporates texture-based feature extraction using the Gray,Level Co-occurrence Matrix (GLCM) for comparative analysis with a Naïve Bayes classifier. The methodology includes comprehensive preprocessing techniques—such as resizing, noise reduction, contrast enhancement, and augmentation (e.g., image flipping and rotation)—to enhance data quality and model generalization. TrueScan also integrates a Streamlit-based web interface for seamless user interaction, enabling clinicians to upload images, adjust visualization parameters, and receive predictions in real time. By automating critical steps in the diagnostic workflow, the system offers improved accuracy, reduced diagnostic burden, and an accessible, cost-effective tool for healthcare practitioners, particularly in resource-limited environments.