Breast density classification using histogram moments of multiple resolution mammograms

Breast density is a strong indicator for breast cancer, which can be assessed by experienced radiologists using mammograms. In this paper, an automatic approach for breast density classification is studied. Mammographic images are pre-processed to separate breast tissues from the background using intensity and morphology-based algorithms. Histograms of multiple resolution mammograms are calculated on the processed images. The statistical moments are retrieved from the multiple resolution histograms, which are employed as the breast density features. The support vector machine (SVM) techniques are implemented onto the feature space to classify the mammograms into different density categories. Experiments on a public dataset verify the performance of the proposed method.