A Computer Vision Model for Early Detection and Classification of Breast Cancer Using Deep Learning Algorithms

Abstract

Breast cancer remains a leading cause of global mortality, though early intervention significantly improves prognoses. The conventional screening technique, X-ray mammography, presents challenges for the early identification of lesions. The compression involved in the imaging process often obscures subtle abnormalities within dense breast tissue. Furthermore, the considerable inter- and intra-patient variability of breast anatomy complicates accurate diagnosis when relying on manually engineered features. Deep learning, a branch of machine learning demanding substantial computational resources, has demonstrated remarkable efficacy in complex, intelligence-driven tasks. This paper introduces a novel neural network architecture, derived from the U-net model, designed for the effective and early detection of breast cancer. The results, which show high sensitivity and specificity, suggest the model's potential utility in clinical environments.