Filtering by approximated densities applied to texture modelling for mammography

Many techniques are currently used for breast abnormality location and breast cancer detection, in particular. One find statistical approaches involving second and/or higher order statistics. The applicability of the filtering by approximated densities (FAD) is here demonstrated. The FAD introduced to alleviate limitations due conventional Kalman modelling, is applied to texture modelling for mammography. This application uses the simplest form of FAD involving second order statistics.