does anyone understand why the center of the image is gray initially? Is it just that the initial guesses are not uniformly random (i.e. farther from the center, we're more likely to guess farther away pixels), and we glossed over that in lecture?
yueli96
I am also confused about this visualization image. If we take random initialization, why pixels in the bottom right corner are more likely to match with pixels in the top left?
ntm
I'm not sure if the center of the image has significantly more gray pixels than other regions; this might be a perceptual factor due to the color-mixing happening in the middle. It seems that the match offset is varying quite a bit in all regions since that's due to saturation.
does anyone understand why the center of the image is gray initially? Is it just that the initial guesses are not uniformly random (i.e. farther from the center, we're more likely to guess farther away pixels), and we glossed over that in lecture?
I am also confused about this visualization image. If we take random initialization, why pixels in the bottom right corner are more likely to match with pixels in the top left?
I'm not sure if the center of the image has significantly more gray pixels than other regions; this might be a perceptual factor due to the color-mixing happening in the middle. It seems that the match offset is varying quite a bit in all regions since that's due to saturation.