Measurements

If we can actually identify an object by some segmentation process involving boundaries, either of brightness or colour or texture, then again we can use mask methods or moments to try to obtain information about the shape of the object in the three colours. This gives us three separate sets of mask values or moments, and they may correlate. Indeed it is to be hoped that they do.

Finding objects out of correlations or textures alone is at the present time a convenience and tends to be image specific, as with the girly pix, rather than a pressing need. This may change

as the objects to be recognised get more complicated and more interesting.