ABSTRACT
In image processing, filters are mainly used to suppress either the high frequencies in the image, that is, smoothing the image, or the low frequencies, that is, enhancing or detecting edges in the image. An image can be ltered either in the frequency or in the spatial domain. In both cases, the intent is to replace some of the pixels with approximate pixels that can improve the overall visualization of the image. In the case of spatial domain ltering, the ltering operation is performed directly on the image without any transformation. Alternatively, in the frequency domain, the image is rst transformed into the frequency domain using FFT-type algorithms and then appropriate ltering is applied. e spatial domain lters will usually rely on operations such as convolution, pixelby-pixel multiplication, and summations, whereas the frequency domain processing requires masking, elimination, and preltering equalization type of operations.
