To expand on the title, let’s say I have a 3 channels image with 5*5
window. I would like to return the coordinates of the max pixel using the sum of all channels using 5*5
window in context of fill block.
A short code from my head that extracts one set of coordinate.
sp cat
+compose_channels +
+z. 0,0,4,4
max_x={xM#-1}
max_y={yM#-1}
That, but in context of fill block, and return rgb using the original image using max_v per crop coordinates.