A comparison of breast tissue classification techniques Arnau Oliver 1 , Jordi Freixenet 1 , Robert Mart´ ı 1 , and Reyer Zwiggelaar 2 1 Institute of Informatics and Applications, University of Girona Campus Montilivi, Ed. P-IV, 17071, Girona, Spain {aoliver,jordif,marly}@eia.udg.es 2 Department of Computer Science, University of Wales Aberystwyth SY23 3DB, UK rrz@uw.ac.uk Abstract. It is widely accepted in the medical community that breast tissue density is an important risk factor for the development of breast cancer. Thus, the development of reliable automatic methods for classifi- cation of breast tissue is justified and necessary. Although different works in this area have been proposed in recent years, only a few are based on the BIRADS classification standard. In this paper we review differ- ent strategies for extracting features in tissue classification systems, and demonstrate, not only the feasibility of estimating breast density using automatic computer vision techniques, but also the benefits of segmen- tation of the breast based on internal tissue information. The evaluation of the methods is based on the full MIAS database classified according to BIRADS categories, and agreement between automatic and manual classification of 82% was obtained. 1 Introduction Breast cancer is considered a major health problem in western countries, and constitutes the most common cancer among women in the European Union. It is calculated that between one in eight and one in twelve women will develop breast cancer during their lifetime [1]. Mammography is still the preferred and most efficient method for detecting breast cancer at early stages, a crucial issue for a high survival rate. Computer Aided Diagnosis (CAD) systems are being developed to assist radiologists in the evaluation of mammographic images [2,3]. However, recent studies have shown that the sensitivity of these systems is significantly decreased as the density of the breast increases while the specificity of the systems remains relatively constant [4]. From a medical point of view, these studies are disap- pointing, because it is well-known that there is a strong positive correlation between breast parenchymal density in mammograms and breast cancer risk [5]. Therefore, automatic classification of breast tissue will be beneficial, not only to estimate the density of the breast, but also to establish an optimal strategy to follow if, for example, the user is looking for masses.