Quantum-Enhanced Cycle Generative Adversarial Network for Image-to-Image Translation
Jari Fathi El, Suchada Sitjongsataporn
This work investigates the integration of quantum computing principles into generative adversarial networks (GANs) for image-to-image translation tasks. A hybrid Quantum CycleGAN architecture was implemented by replacing one convolutional layer in the classical CycleGAN generator with a parameterized quantum circuit, forming a quantum bottleneck block. Both classical and hybrid models were trained on the Monet-to-Photo dataset to evaluate generative quality and computational efficiency. Quantitative metrics, including Fréchet Inception Distance (FID), Kernel Inception Distance (KID), and mean epoch time, were used for comparison.
International Electrical Engineering Congress (iEECON)•Mar 1, 2026•0 citations