Faster subimage extraction?

[dproc]

I want to create a new image composed of a small rectangular region from an existing image. I can copy all the pixels of the existing image into a new image and then crop the new image, but I wonder if there is a way to avoid copying all those pixels (I need speed).

I'm working with Magick++ but any command-line ideas will be helpful because they might show me how to do it in Magick++.

And if there's no way to do this, might I suggest a new Image constructor that takes a Geometry describing the source sub-image?

All thoughts, ramblings, help will be appreciated.

[anthony]

If you need to preserve the original image in memory, why not clone it first, then crop it.