Previous | Next --- Slide 41 of 79
Back to Lecture Thumbnails

How do you decompose an image into frequencies?


Could you clarify/re-explain what is consider high/low frequency aspects of images?


Does this frequency representation give you a complete picture of the image, or do you need to consider some kind of offset as well? For example, intuitively I assume that a solid black image would have 0 frequency, but how do you distinguish that from a solid white image in terms of frequencies?


Can you explain a bit more about the "spectrum" diagram? Is the center of the spectrum image the high frequencies and the more outward you go from the center are the low frequencies?


Could you give a definition of frequency in images? Also, I don't understand why the "spectrum" picture is only black-and-white? Shouldn't a spectrum consists of all colors?


The spectrum diagram is still a bit confusing to me. Also, what does "spatial domain result" mean?


Is frequency conversion invertible? Can we fully recover the original image given the image on the right or does the image on the right require very high resolution to be fully recoverable?


Why does the spectrum diagram have the white vertical and horizontal lines? I can see how the high amplitudes would be gathered in the middle and gradually pan out in a circular way, but the dividing lines in the middle don't seem like they would make much of a difference. Also, does the center gradient fan all the way out to the edge or does it have a smaller diameter?


While I visually seem to understand how the spectrum diagrams for images work with the following examples for different frequency passes, I'm still not really understanding how the spectrum diagram is generated/corresponds to the image. I think it would be helpful to see some examples of the frequency diagram for a few different images or just have some extra explanation here.


How can an image be decomposed into these "frequencies"? What is considered a low or high frequency in an image and how do we compare images on this property?


Why this spectrum diagram is centrosymmetric?


How is a color image decomposed into frequencies? Is the spectrum for each color just taken separately or is there a better way of doing it?


How can we interpret image frequency? The more complex the image is, the higher frequency it has?


I'm a bit confused about the spectrum diagram. Why is the light centered?


Is image decomposition very efficient in CPU?


How is the frequency of an image defined? Lightwave?


How do we calculate the spectrum for an image?