A Deep Unsupervised Learning Method for Degrading Mammogram Restoration Scheme

Abstract

Recent trend medical image analysis plays a crucial role for the diagnosis of different diseases like breast cancer, lung cancer, liver cancer, brain tumor, etc. Image denoising is a significant preprocessing step during the processing of medical images. Many vital structures are not perceptible properly since medical images are poorly illuminated. Lots of uncertainties are present in the image in imprecise form, which leads to inaccurate diagnosis. Thus, it is essential to remove noise from the image or to restore the image. The objective of denoising of an image is to renovate an image to another form that is more appropriate for further processing. A number of image denoising algorithms have been established in the past few decades. Recently, deep learning based image denoising scheme has presented outstanding performance compared to other straight image denoising methods. In this exploration, we present an image denoising strategy in the light of a convolutional denoising autoencoder and assess clinical applications by looking at existing image denoising methods.