Mammogram Denoising Using High Boost Filter and Vectorization Convolutional Neural Networks

Abstract

Mammogram screening is the effective way to prevent breast cancer in the early detection stage. However, low dose radiation from x-ray machine causes a degradation of visuality of mammograms that make interpretation error to radiologists. In this paper, high boost filter and vectorization convolutional neural networks are combined to improve visual quality of mammograms. The experimental results illustrate that the proposed method can improve contrast and suppress noise in mammograms comparing to the traditional denoising methods.