Computer Application for Enhanced Breast Cancer Detection Based on Extreme Machine Learning Algorithms

Breast cancer remains one of the leading causes of dead among women worldwide, highlighting the importance of early and accurate detection to enhance treatment outcomes and survival rates. This study introduces a computational application developed for advanced breast cancer detection utilizing Extreme Learning Machine (ELM) algorithms on histological images. The dataset employed, BreakHis, consists of 1,807 samples magnified at 400X, facilitating an in-depth analysis of tissue characteristics. The investigation evaluates both standard and regularized variants of the ELM algorithm to identify the most efficient model in terms of performance and computational cost. Additionally, the Scaled Conjugate Gradient (SCG) method is utilized for comparative analysis against ELM outcomes. Performance metrics, including accuracy (ACC), geometric mean (G-mean), confusion matrix, and time complexity, are collected to provide a comprehensive evaluation of the models. Ultimately, an application has been developed to validate diagnoses through the upload of histological images and corresponding prior diagnoses, with the goal of implementation in clinical and research settings.