Introduction#

Filtering in image processing is a very precise notion that differs from the filters proposed by some apps or softwares. In image processing, filtering consists in amplifying or attenuating some frequencies in the image. To study this notion, we need to introduce two mathematical objects, namely the Convolution and the Fourier transform, thus we will consider both the usual spatial domain of the images, but also the frequency domain. The Fourier transform is the tool to pass from a domain to the other. To know more about the Fourier transform in signal processing, you can read this very excellent course (in French).