Effect of matrix size in affecting noise reduction level of filtering

Abstract

Improved image quality needs to be done to improve data processing on the image. This quality improvement can be made by doing masking technique. Median Filter is one technique to enhance the quality of an image in a particular space. Median filtering improves the image by specifying a specific pixel from its neighboring pixels. The median filtering calculation uses a matrix block with an odd number. Each matrix block will have a middle value after the pixel values have regularly been sorted. This method is included in the category of nonlinear filtering. With the median filtering, the output pixel value is determined by the median of the specified mask environment. Median Filter has different results when using different matrix sizes as well. The results of this process can determine how gentle the result of noise reduction. In general, the larger the size of the matrix, the higher the blurriness of an image.

Experimental Study of Modified Multilevel Median Filter for Noise Reduction

“…Besides being able to suppress noise, the median filter can also smooth the image according to the spatial window size used [10][11] Figure 1. Sub-windows used by Multilevel Filters [12].…”