Computer Science

Quantum Deep Generative Prior with Programmable Quantum Circuits

T. Xiao, X. Zhai, et al.

Discover the groundbreaking Quantum Deep Generative Prior (QDGP) algorithm developed by Tailong Xiao, Xinliang Zhai, Jingzheng Huang, Jianping Fan, and Guihua Zeng. This innovative approach utilizes programmable quantum circuits to significantly enhance image reconstruction and generalization in computer vision tasks, surpassing classical methods, especially at low sampling rates. Dive into the quantum revolution in imaging!

00:00

Playback language: English

Index

Abstract

This paper proposes a quantum-enhanced deep generative algorithm, Quantum Deep Generative Prior (QDGP), using programmable quantum circuits to generate quantum latent codes. The algorithm demonstrates superior reconstruction performance in optical ghost imaging experiments compared to classical methods, particularly at lower sampling rates. It also shows enhanced generalization capabilities in various computer vision tasks like image inpainting and colorization, exceeding the performance of classical counterparts.

Publisher

Communications Physics

Published On

Aug 15, 2024

Authors

Tailong Xiao, Xinliang Zhai, Jingzheng Huang, Jianping Fan, Guihua Zeng