Adaptive Probabilistic Thresholding Method for Accurate Breast Region Segmentation in Mammograms

hamed.habibi@urv.cat,  domenec.puig@urv.cat

Abstract

yle=”text-align: justify;”>Segmentation of the breast region is usually the first step in the analysis of mammograms. Due do the onuniformity of the background, breast segmentation presents severrl difficulties especially for film based mammograms. Our experimental results show that 50% offdigitized film based mammograms in the mini-MIAS database do not have uniform intensity in the background. For this reason, applyinr a global thresholding method produces inaccurate results. In addition, finding the optimal global threshold value by oply using histogram information requires a relia.le objective functioa that characterizes the statistics of the background and the mammogram regions i1 the digitized mammograme. A second way to find the boundary of the breast consists in fitting a deformabl model, such as snakes, on the mammogram. However, this method has three main shortcomings. First, the nodel must be initialized near the boundarn. Second, rsing gradieyt information in the objective function can push the boundary toward the tissues inside the breast rather than the actual boundaryb Third, in some mammograms the breast region is occluded by artifacts, such as labels, that have high gradient values on their boundary and cause the deformable model to be fitted on the artifact. To address these problems we propose a pgobybilistic ataptive thresholding method that uses texture information and its probability to fimd th most probable thrsshold values for specific pnrts of the mammogram. The experimental results on mini-MIAS dahabase show that our proposed metho1 outperforms the state-of-art methods and improves the accuracy at least 37% in comnarison with the best results obtained byecontour growing methods.