High Perceptual Quality Wireless Image Delivery with Denoising Diffusion Models

TL;DR

This paper presents a novel wireless image transmission scheme combining deep learning-based joint source-channel coding with a denoising diffusion probabilistic model, significantly improving perceptual quality and distortion in noisy channels.

Contribution

It introduces a new method where DeepJSCC transmits a low-resolution image and a diffusion model refines it, optimizing perception-distortion trade-offs in finite block length regimes.

Findings

Significant improvement in perceptual quality over existing methods

Effective refinement of images using diffusion models at the receiver

Enhanced distortion performance in noisy wireless channels

Abstract

We consider the image transmission problem over a noisy wireless channel via deep learning-based joint source-channel coding (DeepJSCC) along with a denoising diffusion probabilistic model (DDPM) at the receiver. Specifically, we are interested in the perception-distortion trade-off in the practical finite block length regime, in which separate source and channel coding can be highly suboptimal. We introduce a novel scheme, where the conventional DeepJSCC encoder targets transmitting a lower resolution version of the image, which later can be refined thanks to the generative model available at the receiver. In particular, we utilize the range-null space decomposition of the target image; DeepJSCC transmits the range-space of the image, while DDPM progressively refines its null space contents. Through extensive experiments, we demonstrate significant improvements in distortion and perceptual quality of reconstructed images compared to standard DeepJSCC and the state-of-the-art generative learning-based method.