Identification of most preferential denoising method for mammogram images

Abstract

Mammogram plays a vital role in clinical imaging. It is necessary to provide a clear image to the surgeon for diagnosing the disorder and diseases of soft and complex tissue structure. Images obtained from the mammogram may have noises added to it during capturing of the image. Removing noise is still a challenging problem. Many filters are innovated to remove noise from the image with its postulation, advantage, and limitations. In this paper, Hybrid Median Filter (HMF), Linear Minimum Mean-Square-Error (LMMSE), Oriented Rician Noise Reduction Anisotropic Diffusion (ORNRAD), Higher Order Hybrid Median (HOHM), and Non-Level Means (NLM) denoising filters are used to remove noise from a mammogram image. The performance analyses of filters were evaluated by Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Image Quality Index (IQI), Mean Absolute Error (MAE), and Contrast to Noise Ratio (CNR) parameters. ORNRAD filter gives desirable results regarding above five quality and performance analysis attributes.