A Quantitative Study of Local Ternary Patterns for Risk Assessment in Mammography

This paper presents a preliminary quantitative study for breast cancer risk assessment in mammography using mathematical operators called Local Ternary Patterns. The study covers three different mapping patterns namely uniform ('u2'), nonuniform (`'ri') and a combination of uniform and nonuniform ('riu2'). These patterns are used as texture features to model the appearance of breast density within the fibroglandular disk area. Subsequently, the Support Vector Machine is employed as a classification approach and initial results suggest that the mapping pattern `'riu2' outperforms the others.