Compressed sensing is an image/signal processing algorithm that allows you to re-construct an image/signal even when you’ve lost up to 95% of the samples. It’s so good that it can even be cranked up to restore images even above what would normally be the Nyquist limit.

Applied Science walks through using an X-ray backscatter device to reconstruct images as near to x-ray vision as you can get at low doses.

Links:

• Robert Taylor paper on Compression Sensing in Python.
• Steve Brunton’s introduction to compressed sensing