The code I sent does seem to work, but the mouse input is difficult
to get right. By default it uses image coordinates, while the CI
kernel really wants pixels. So I don't know the best way to line
up the mouse input with the image, but I've put together a little
composition that works except that the mouse position is offset
from the image. It uses a video input, but if you don't have a
video source replace it with an Image Importer.
Use Rendering Destination dimensions to get the ratio between units
and pixels (R) along with the coordinate system width (W) and height
(H).
X' = (W/2 + X) * R
Y' = (H/2 + Y) * R
This formula can be done using a single Math patch.
Use a Round patch if you want to round the closest integer pixel
coordinate.
