Breast Cancer Diagnosis with Histopathology Images using Adaptive Feature Extraction and Machine Learning Model

Histopathology photos provide essential data for researchers investigating cancer biology and formulating innovative tactics and analysis. Through the analysis of extensive databases of histopathological pictures, researchers can identify biomarkers, comprehend their classification patterns, and uncover prospective targets for pharmacological development. This research use a combination of feature extraction techniques and machine learning algorithms to categorize breast histopathology pictures. The uneven dataset is rendered balanced through augmentation. A sequence of preparation steps, including grayscale image conversion and picture sharpening using a Wiener filter, is implemented. The local binary patterns (LBP) extract features from photos. Machine learning models such as logistic regression, linear discriminant analysis (LDA), random forest, XGBoost, and support vector machine (SVM) utilize the retrieved features. Logistic regression surpasses the other classifiers, achieving an accuracy of 98.88.