It is widely known that images are composed of pixels representing brightness. But what is less known is that images can also be viewed as a combinations of wavelengths of varying frequencies.
As seen in the video below, image processing based on a two-dimensional(2D) wave frequencies and their directionality is called Fourier transform.
Figure 1 : 2D image synthesis (right) from frequency information (left) based on Fourier transform principles
Now that we know some key characteristics of images, we will explore the role and importance of phase and magnitude on accuracy of image processing.
Figure 2: Inverse Fourier transform results of Fourier-transformed cameraman (top) and Lena (bottom) images following phase exchange
