On this day in 1919, Godfrey Hounsfield was born. Here is the mechanism at kitchen-table size. A normal X-ray is a shadow: everything the beam passes through gets flattened onto one plate, so the loaf and the coin inside it overlap. Hounsfield's insight, shared in its mathematics with Allan Cormack, was that if you shine the beam through the loaf from many angles and simply measure how much comes out the other side each time, you have a pile of sums with one solution. Arithmetic can work backwards and say how dense each crumb must have been. Not a shadow, then, but a slice.
What I love is the beginning of it. The first rig was built on a lathe bed, turning the specimen a degree at a time, and the specimens included a preserved human brain. The early scans took hours to gather and longer still to compute, which is exactly the sort of patience my own lab days lacked. By 1971 a version of it was scanning patients, and in 1979 he and Cormack shared the Nobel Prize in Physiology or Medicine.
Radiologists still measure tissue density in Hounsfield units, which seems a fair memorial for an engineer who wanted to know what was inside without opening it. The Wikipedia article is linked below.
Ada, here's the clever bit of the funding: EMI's research pocket in those years was kept comfortably full by record sales, so a good deal of the first scanner was paid for by people buying singles. The reconstruction sums took hours on the early machines — the patience was in the arithmetic as much as the lathe bed.
Ada, the loveliest part for me is that the same reconstruction maths turns up in my field too: radio astronomers use it to build images of the sky, and spacecraft have used it to sound the inside of a comet's nucleus. One man's loaf of bread is another's dirty snowball.