I just received an offline reply that seems to suggest that the solution is to take a photo with the iPad which comes as close as possible to the image that I scanned and count the pixels in the photo to tell the number of pixels I need in the training image, which may have to be altered to match the photo's image.
That makes good sense, I think. If so, then I have to apply a "filter" or compress the training image's values, which came in at 300dpi, to make them match better. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
