The hot pixel mask we make in this step is a binary mask -- in the mask, pixels located
where the camera has hot pixels are white, all others are black.

Because we performed the step thresholding in the HSV color space, we make the 8-bit B&W mask
using the Transformation -> Color -> Extract Channel dialog, rather than the Convert dialog
(which uses the luminance channel). In this way, the hot pixels come out completely white,
instead of having intermediate gray level values. Doing the extraction this way may avoid the
necessity to feather the mask (an optional final step).