The focus of this paper is to review approaches for segmentation of breast regions in mammograms according to breast density. Studies based on density have been undertaken because of the relationship between breast cancer and density. Breast cancer usually occurs in the fibroglandular area of breast tissue, which appears bright on mammograms and is described as breast density. Most of the studies are focused on the classification methods for glandular tissue detection. Others highlighted on the segmentation methods for fibroglandular tissue, while few researchers performed segmentation of the breast anatomical regions based on density. There have also been works on the segmentation of other specific parts of breast regions such as either detection of nipple position, skin-air interface or pectoral muscles. The problems on the evaluation performance of the segmentation results in relation to ground truth are also discussed in this paper.