6. Image extension

6.1 Basic principles

All the digital linear filters discussed in this article for 1D (one-dimensional) signals are applicable in the case of images (two-dimensional signals). However, while the notion of low or high frequency is natural for 1D signals (e.g. a low-pitched sound or a high-pitched sound), this is not the case for images. Indeed, for images, it's a question of spatial frequency: details frequently repeated over a small number of pixels correspond to a high frequency (fine textures, contours...), while small variations diluted over large areas of the image (background variation of an image) correspond to low frequencies. For these reasons, image filtering will not be specified on a frequency basis.

In this way, image filtering can be used to modify...