Year: 2026
Type: article
Abstract: Ensuring both perceptual quality and semantic fidelity in image transmission remains a critical challenge for communication systems, as conventional methods often suffer from bandwidth limitations, noise sensitivity, and inadequate preservation of contextual information. To address these issues, this study proposes a quantum frequency-domain framework that jointly performs image compression and transmission while enabling semantic-aware recovery using a novel channel-aware denoising framework. The process begins by converting image bitstreams, which may be optionally channel encoded, into quantum states using variable-size qubit encodings. The Quantum Fourier Transform (QFT) is then applied to map these quantum states into the frequency domain, where compression is achieved by transmitting only a minimal subset of frequency components. At the receiver, the corresponding decoding processes are performed. Finally, the reconstructed image is refined using a channel-aware diffusion-based denoising framework, which enhances perceptual quality by restoring fine details and preserving semantic information lost due to noise and compression. Experimental results, evaluated using objective quality analysis, show that higher qubit encodings improve compression efficiency and reduce channel capacity requirements but are more vulnerable to noise. To address this trade-off, the proposed channel-aware diffusion-based denoising framework effectively enhances semantic fidelity, leading to improved noise resilience and superior reconstruction quality, even under severely degraded channel conditions.
Institution: University of Strathclyde
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7203698463