MGRegBench: A Novel Benchmark Dataset with Anatomical Landmarks for Mammography Image Registration
Researchers have introduced MGRegBench, a new benchmark dataset and evaluation protocol designed to standardize and improve mammography image registration. This public dataset includes over 5,000 image pairs, segmentation masks, and manually annotated anatomical landmarks, along with baseline implementations. MGRegBench aims to enable fair, reproducible comparisons of various registration methods, from classical to deep learning approaches, and foster further research in AI-driven medical imaging. AI
IMPACT Establishes a standardized benchmark for AI-driven mammography registration, enabling better comparison and development of medical imaging tools.