TAF PUBLISHING

The aim of this study is to harness the great potential of image processing techniques which have evolved significantly in the last years, to build an automatic system to detect and diagnose breast cancer in the digital mammographic images in order to help those interested people in this domain, such as radiologists and specialists in oncology and to improve their performance by reducing error rates of breast cancer diagnosis. As long as segmentation and extracting the effective features of mammograms play a major role to isolate and classify suspicious regions which can be subject to cancer, in this work, we focus on abnormality detection using multi-thresholding OTSU's method to segment the Region Of Interest (ROI). Then the texture features of the segmented ROI are extracted which are used to classify the ROI as normal or abnormal tissue by using an Artificial Neural Network (ANN). This system can correctly classify the tested region by a rate of 93.80%.