Sorry to get you confused. This is a different setup. TFA uses the 30x30 and no compressed sensing. The link above uses a single photo detector. They also use an LED matrix, but that's to make the _image_ (I think)