The Image > Mode > 32 Bit Complex command converts the active image into an FT image.
Prerequisite: This command is only available when the active document is a 16-bit gray-value image.
What is a Fourier transform?
Every periodic function can be presented as a sum of harmonic functions. The Fourier transform (FT) breaks down the source function into sine functions of varying frequencies. Images located in the spatial domain are transformed into the frequency domain by the Fourier transform.
The image on the left displays a gold sample as viewed by a transmission electron microscope. The image shows periodic structures caused by the gold crystal lattice. On the right, you can see the FT image. Every point in the FT image represents a particular spatial frequency. The gold's lattice spacing determines the distance of the spots from the center of the FT image.
What happens to the image information?
The FT image contains the same image information as the source image. It's just displayed differently in the FT image.
This means that you can convert an FT image back into a 16-bit true-color image. To do so, use the Image > Mode > 16 Bit / Channel command. This conversion is an inverted Fourier transform. The resulting image looks like the source image. Alternatively, use the Inverse FT button in the Fourier transform tool window's toolbar to carry out an inverse Fourier transform. You can find more information on this topic here.
Note: It's possible that when an image is transformed back again using the Image > Mode > 16 Bit / Channel command, that it will be appear to have very little contrast. This could be because it's not being displayed optimally on the monitor. Show the Adjust Display tool window. Select the Auto Contrast option and click the Apply button.
On the left you see the FT image and on the right the corresponding 16-bit gray value image. Both images contain the same image information. The only difference is that in the FT image you see the spatial frequency distribution and in the 16-bit gray-value image you see the sample in the spatial domain, as usual.
See also
5246 14032014