Filtering Operation on Images

The methods described till this point were working on a single pixel intensity which is called as point processing methods. Sometimes it is helpful to look at a neighborhood of pixel rather than only single pixel intensity. It is called as neighborhood processing techniques. The neighborhood can be 3x3, 5x5, 7x7 etc matrix centered at a particular pixel. Image filtering is an important neighborhood processing technique. 

Filtering is an important concept in signal processing where we reject a certain band of frequencies and allow a certain band of frequency to pass. how frequency is measured in Images? If gray levels changes slowly over a region then it is a low-frequency region. If grey levels changes drastically then it is a high-frequency region.  Normally background of an image is considered as low-frequency region and edges are high-frequency region. Convolution is a very important mathematical concept for neighborhood processing and Image filtering in particular...