Classification of Mammogram Patterns using area measurements and the Standard Mammogram Form (SMF)

Our aim is to propose a new approach to breast pattern classification that will aid in the development of an automated mammographic density analysis procedure. Breast patterns broadly classify the mammographic density and density distribution of each mammogram in order to provide a framework for assessing the risk of breast cancer according to density. However, basic weaknesses of classifying mammogram films are the inter and intra-subject variability in interpretation and coding of the patterns by different radiologists as well as the fact that the intensities of the mammogram do not indicate the actual volume of the breast regions. Our endeavour is the development of objective descriptors of breast density based on the digital mammogram. Unlike previous efforts to classify breast patterns based on image texture, we propose the use of density measures based on a normalised representation of the breast (hint representation or SMF) and demonstrate that they can provide a more intuitive quantitative description of mammogram patterns.