Friday, January 16, 2009

Compressive/Compressed Sensing/Sampling

So which is it?

Let me suggest that "Compressive" is the term we want to use rather than "Compressed". If something is already compressed, why should we bother with it unless we wanted to lift it back into its ambient space? When we sense or sample, we want to reduce dimensionality, thus "Compressive" is a more proper term.

Secondly, there should be a distinction between "Sampling" and "Sensing". When we sense, it is from an environment that is irreversibly changing. When we sample, it is from a population that is for the purposes of the experiment, static in some respect. So, "Compressive Sensing" is what we do when we convert measurements from a dynamic environment to information. "Compressive Sampling" is what we do when we down convert a statics set such as a previously captured image ordered set of pixels.