FQGA-single: Towards Fewer Training Epochs and Fewer Model Parameters for Image-to-Image Translation Tasks

TL;DR

This paper introduces FQGA-single, a new model for medical image translation that achieves high-quality results with fewer training epochs and parameters, outperforming existing models like CycleGAN.

Contribution

FQGA-single is a novel model that significantly reduces training time and model complexity while surpassing CycleGAN in medical image translation quality.

Findings

FQGA-single trained on a single epoch outperforms multi-epoch training.

FQGA-single surpasses CycleGAN in quality with fewer parameters.

Single-epoch training yields better results than multiple epochs for FQGA-single.

Abstract

This paper proposes a novel model inspired by CycleGAN: FQGA-single to produce high quality medical synthetic CT (sCT) generated images more efficiently. Evaluations were done on the SynthRAD Grand Challenge dataset with the CycleGAN model used for benchmarking and for comparing the quality of CBCT-to-sCT generated images from both a quantitative and qualitative perspective. Finally, this paper also explores ideas from the paper "One Epoch Is All You Need" to compare models trained on a single epoch versus multiple epochs. Astonishing results from FQGA-single were obtained during this exploratory experiment, which show that the performance of FQGA-single when trained on a single epoch surpasses itself when trained on multiple epochs. More surprising is that its performance also surpasses CycleGAN's multiple-epoch and single-epoch models, and even a modified version of CycleGAN.