Abstract

GUERRA, Aníbal  and  RIVAS, Joel. Micro-calcification detection in mammographic images using neural networks. Rev. Fac. Ing. UCV [online]. 2011, vol.26, n.3, pp.7-14. ISSN 0798-4065.

According to the American Cancer Society, breast cancer is one of the leading causes of death in women worldwide. The key factor to reduce the impact of this disease is an early diagnosis. The software described in this document aims to be a mechanism for second opinion in detection of micro-calcifications in mammographic images. In this software, digitalized mammographies are processed and inserted as entry data to a multi-layer perceptron, which is able to detect presence of micro-calcifications in the provided images. The neural network was implemented in C++ language; its architecture has one hidden layer and uses a back-propagation learning algorithm in combination with techniques of statistical analysis over the image texture. The data used in this research was extracted from the MIAS database. The software assessment reported 94.4% of sensitivity in prediction tasks, showing the potential of both techniques in the resolution of the problem.

Keywords : Mammography; Artificial neural network; Computer aided detection; Micro-calcification.