Score: 0 / 4
. Why does the lesson prefer Y = 0.299 R + 0.587 G + 0.114 B over a plain average (R + G + B)/3 for grayscale conversion?
A plain average can't tell green and blue apart perceptually; weighting toward green (as human vision does) gives a much better brightness estimate.
. In HSV, which channel should you threshold on to segment an object by color robustly, even under a lighting gradient across its surface?
Hue captures which color something is, independent of how bright or washed-out it appears, so it stays consistent even as brightness varies across the object.
. Why does video/JPEG compression discard resolution from the chroma channels (Cb, Cr) more aggressively than from luma (Y)?
Chroma subsampling exploits the fact that losing the same amount of detail is far less objectionable in color than in brightness.
. Why was the L*a*b* color space designed the way it was?
Equal RGB distances can correspond to wildly different perceived differences depending on where they fall in color space; Lab was designed so Euclidean distance tracks perception much more consistently.