Fourier transformation can be used to get the underlying frequencies out of a time function. **Discrete Fourier transformation** or **DFT**, on the other hand, is a method to calculate the underlying frequencies of a sampled time function (thus discrete). That is a purely mathematical definition and a very short one in that sense, so to put it in terms of computer vision and image processing, you need to first try to think about an image (a grayscale image) as a discrete distribution of points on a three dimensional space, where `X` and `Y` of each discrete element are the pixel position in the image and `Z` is the intensity value of the pixel. If you are able to do that, then you can also imagine that there is such a function ...