The aim of image smoothing is to diminish the effects of camera
noise, spurious pixel values, missing pixel values, etc. It is commonly
achieved using neighborhood averaging and edge-preserving
Each pixel of the smoothed output image, F'(x, y) is obtained from the average value of the neighborhood of the respective (x, y) in the input image. For instance, if a 3×3 window, as shown in Fig. is used,
the intensities of the pixels covered by the window are controlled by 1/9
and summed. The resultant value is used to replace the intensity of the central pixel in the window.
This mask is successively moved across the image until every pixel is processed. Gaussian blur is yet another commonly used technique for image smoothing. Hence, smoothing reduces or attenuates the higher frequencies in the image.
Edge Preserving Smoothing
Neighborhood averaging and Gaussian smoothing operations tend to blur the edges of an image as the high frequencies are attenuated in the output image. An alternative approach is to use median filtering,
which sets the intensity value of each pixel to the median of its neighborhood pixels. This technique assures significant edge preservation.
The smoothing effect of neighborhood averaging (mean filter)
and edge-preserving smoothing (median filter) is shown in Fig.