Auto-probing breast cancer mass segmentation for early detection

In this paper, an auto-probing breast cancer mass segmentation (ABC-MS) is proposed to assist medical doctors in breast cancer diagnosis. Manual segmentation is implemented as standard diagnosis procedure for medical doctors. This algorithm can detect and segment the breast cancer abnormality without prior knowledge regarding its presence. Automated single seed point region growing is utilised in this algorithm to perform the mass detection and segmentation automatically. Comparison with commercial semi-automated segmentation application is performed. Tabulated experiment results showed the proposed method outperformed the compared method by having accuracy at 90% and area under curve (AUC) at 0.895 ± 0.0338.