A 2-phase U-Net trained on abnormality-focused patches achieved a mean Dice coefficient of 0.58 and detected 77.8% of mammographic lesions on a 55-image test set, with 100% detection of large lesions but resolution-dependent failures for small abnormalities.