Image data as a sine wave?

Does anybody know if it is possible to represent image data as a sine wave?

Answer 1

[Image Analyst]

More info is needed. I can represent an image by a lot of sine waves - that's what its Fourier transform is.

[Philip]

I have a 1 dimensional image vector, where the centre element corresponds to a strong edge. What I want to do is show that compressed images exhibit the ringing artifact (i.e. overshoot and periodic ups and downs), whilst the same properties do not exist for uncompressed images. As you can tell, my maths is not great, but I was hoping to illustrate this by producing an estimate of a clean sine-wave for this signal, against that of what is shown when the image is compressed/uncompressed. This would then enable me to calculate the size of the overshoot and harmonics associated with the ringing artifact... perhaps there is a much better way of doing this, that I am missing?

[Walter Roberson]

fft() would seem to be appropriate. Just watch out for the fact that the output of fft() will very likely be complex.

[Philip]

Thanks, I've been looking at using the output of FFT and I think it's a good way to go. Do you know how I can use the stepinfo() function on the image data, such that it can tell me the overshoot, undershoot and settling time? The document says that the input data must be a Transfer Function, but I am not sure how to compute this from the spatial pixel values...