Rate-Distortion-Perception Theory for the Quadratic Wasserstein Space
This work addresses theoretical limits in information theory for compression and perception, but it is incremental as it extends existing rate-distortion theory to include perception constraints.
The paper establishes a single-letter characterization of the fundamental distortion-rate-perception tradeoff under squared error distortion and squared Wasserstein-2 perception measures, and explicitly evaluates this for Gaussian sources while clarifying universal representation notions.
We establish a single-letter characterization of the fundamental distortion-rate-perception tradeoff with limited common randomness under the squared error distortion measure and the squared Wasserstein-2 perception measure. Moreover, it is shown that this single-letter characterization can be explicitly evaluated for the Gaussian source. Various notions of universal representation are also clarified.