Indian Journal of Science and Technology

Abstract

Background/Objectives: One of the leading causes of death for women is breast cancer, and extensive research has been conducted to improve the diagnosis and detection of breast cancer using various image processing techniques. Medical imaging plays a crucial role in this domain, particularly mammography, which is widely used for breast cancer screening and diagnosis. This paper introduces a novel filtering technique called the New Robust Adaptive Median Filter (RAMF). Method: The suggested approach only takes into account noise-free pixels when determining the window’s median. The median is computed from the remaining pixel values when the ◦ or 255 pixel values are excluded. In order to filter high densities of salt-andpepper noise, the adaptive windowing approach is applied, which enables our algorithm to extend the size of its filtering window dependent on the local noise density. Moreover, a threshold value is employed to establish the pixel value under extreme circumstances, such as pure black and white photos with noise. Finding: To compare filters based metrics, we find evaluate the filters using a standardized dataset and calculate the MSE, PSNR, and UQI values for each filter. These values can then be compared to determine which filter performs better in terms of noise reduction and image quality enhancement. The Proposed Filter shows good performance in low and higher density ranges (10%-90%) to effectively reduce noise in higher density values. Novelty: The New Robust Adaptive Median Filter (RAMF) is a novel filtering technique that aims to reduce noise in images, particularly in the presence of highly corrupted or noisy pixels. This filtering algorithm employs an adaptive approach where the median is calculated in a processing window, but without considering the noisy pixels during the initial computation.

Keywords: Adaptive Median Filter (AMF); Weighted Median Filter (WMF); Noise Adaptive Fuzzy Switching Median (NAFSM) filter; Decision-based algorithm (DBA); Mean Square Error (MSE); Peak-Signal-to-Noise Ratio (PSNR) and Universal Quality Index (UQI)