18- Keeping the Two Gaussians Straight

Finally, a word of warning or clarification at least. We just talked about sigma as being the width of a Gaussian, where that was the variance of the, of the smoo, smoothing of the blurring. Last time we talked about sigma or sigma squared as the variance of a noise function, how much noise was being added. So in one case, where we share might the filter, the sigma is in space, okay? Where as in the noise case it's in an intensity or it’s in the value. The bigger the noise signal is, the more noise you added, the bigger the blurring filter sigma, the more you're going to blur. So, you have to be an in a reasonable called sigma as they're both using a normal distribution. But one is over a space, and one is over intensity. 


We can show those two sigmas here, and by the way, here I'm using I'm using images that go from zero to one. Partially because that I know how sigma varies with respect to the range of the image, and partially because the slides that I stole off the Internet did this. All right? So in the top row, there's no smoothing going on, so we have a sigma of 0.2 in the noise. That's a lot of noise if the range is only going from 0 to 1 or, or let's say minus 0.5 to 0.5. A sigma 0.1 is less noise. A sigma 0.05 is less noise yet. But then we can smooth it, all right, with a Gaussian. Right, this is the smoothing, with the Gaussian kernel. And the more we smooth, the blurrier these things get. And so, for the same amount of smoothing, the thing with the smaller amount of noise, with the same amount of smoothing becomes even smoother, right? So this image here is even smoother than that image there. Okay? But this is showing you those two sigmas, the two Gaussians. Almost always it'll be clear from context, but since I've had situations where students say, wait a minute, I thought the bigger signal was, the more noise we got. Now you're telling me the more sigma is, the more blurring is,the less noise we have, and I say yep, two sigmas.

  

Last modified: Saturday, 6 April 2024, 10:40 AM