Quantum physics is taken like a modern buzz-word, but it is more than a century old. It started off with Max Planck explaining the colour of sunlight (blackbody radiation), and got Einstein his Nobel prize (in 1905) for the photoelectric effect, which is what is used in camera sensors till today.
There are two main differences between classical and quantum theory: the first is that quantum theory, unlike classical physics, is probabilistic, i.e. noise is well understood, comes from fundamental principles (no-cloning) and can be accurately modeled. The amount of information and of noise in a signal can be accurately described from first principles, especially when quantum theory is combined with Communications theory as developed by Claude Shannon (around 1948).
Unlike classical theory, where noise is dominated by your measurement apparatus, in the quantum regime, noise is a property of the signal itself. For maybe 5 or 10 years now, image sensors have been firmly working in this quantum regime, and Rawsie has been developed to work in just that regime. JPEG and the other compression algorithms, had been developed when sensors were still fairly inefficient and limited by their electronics rather than by the quantum nature of light.
Rawsie uses three things that we learned from quantum theory: the way we model noise, the way we characterize the sensors, and the way we estimate and limit the information loss due to compression.
The second difference between quantum and classical is that quantum particles do not obbey “local realism”, i.e. they may be “entangled” meaning that although the properties of two particles are fundamentally random individually, they still share a global property (such as energy, momentum or other). Quantum computers use these aspects to scale performance of some algorithms better than classical ones, quantum computers are still a few years off, and Rawsie does not use that aspect of quantum mechanics.