OK, so I’ve put some alligators through a CT scanner to create three-dimensional models of their cloacas. How exactly is that going to work?
A CT scanner is an ordinary X-ray machine on steroids. The basics for both machines are the same: stick an object-to-be-imaged between an X-ray source and an X-ray detector, and measure the number of X-rays that make it through the object. X-rays are a form of light and do all the same things as ordinary visible wavelengths when they interact with the stuff of the world: some materials bend them, some reflect them, some absorb them. But because X-rays are much more energetic than visible light, more materials are transparent to them. Most visible light bounces off our skin, but to an X-ray it’s as clear as a windowpane.
But an animal’s body isn’t completely transparent to X-rays – molecules in the muscle, fat, and bone can absorb or bend some of the light as it passes through. When this happens, fewer X-rays make it to the detector, and these X-ray-deficient shadows form pictures of the stuff inside the animal.
Your run of the mill X-ray machine takes a picture through a single plane. You’ve no doubt experienced this at the dentist: they stick a detector in your mouth, point a X-ray source at your cheek, and bam! Image of your teeth, roots and all.
CT scanners work on the same principle, but mount the X-ray source and detector in a ring that rotates around the specimen (or person, or whatever) and take many many more pictures in the process. The pictures are fed into a computer program that performs some mathematics I don’t even pretend to understand (this is the “computed tomography” - “CT” - part) and hands back a stack of images that look like a orderly series of planes cut through the specimen. This is a “Z-stack”, and it looks like this:
This particular one (from Julio Pereira) is a stack of images through a human head and chest. Even though the image is in shades of gray, you can clearly see different kinds of tissue: brain, bone, muscle, lungs – all thanks to differences in gray value. Lighter tones represent denser tissues, so bones appear white and the air-filled lungs look black. (Scanning through this particular stack, you can also see a spot where some metal dental work scatters the X-rays into starbursts.)
Importing a Z-stack into specialized image segmentation software makes it possible (again, using some complicated math) to use these grayscale differences to define the surfaces of different tissues throughout the stack and create 3-dimensional models that can be virtually dissected. This fish (scanned and reconstructed by Sarah Faulwetter) is one example: starting with a Z-stack from a CT scan, it's possible to make a model of the surface of the fish that has a model of its skeleton embedded inside it.