Self-Similarity Analysis Applied to 2D Breast Cancer Imaging

This article presents a new trend in computerized medical image analysis of breast cancer features, existing in mammograms (grey scale 2D images), based on recent approaches of the self-similarity formalism. A procedure for detecting small-sized details in mammograms, based on a multifractal approach, is proposed. These details can represent microcalcifications, possible signs of breast cancer. We present and discuss the results obtained using the proposed extraction method. Furthermore, we demonstrate its accuracy when applied to clinical mammograms, revealing microcalcification regions, distinguished by direct singularity information extraction.