Resumen
The appearance of microcalcifications in mammogram images is an essential predictor for radiologists to detect early-stage breast cancer. This study aims to demonstrate the strength of persistent homology (PH) in noise filtering and feature extraction integrated with machine learning models in classifying microcalcifications into benign and malignant cases. The methods are implemented on two public mammography datasets: the Mammographic Image Analysis Society (MIAS) and Digital Database for Screening Mammography (DDSM). This study discovered that PH-based machine learning techniques can improve classification accuracy, which could benefit radiologists and clinicians in early diagnosis.