Evaluating the Performance of StyleGAN2-ADA on Medical Images

McKell K. Woodland, John Wood, Brian M. Anderson, Suprateek Kundu, Ethan Lin, Eugene Koay, Bruno Odisio, Caroline Chung, Hyunseon Christine Kang, Aradhana M. Venkatesan, Sireesha Yedururi, Brian De, Yuan Mao Lin, Ankit B. Patel, Kristy K. Brock

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Although generative adversarial networks (GANs) have shown promise in medical imaging, they have four main limitations that impede their utility: computational cost, data requirements, reliable evaluation measures, and training complexity. Our work investigates each of these obstacles in a novel application of StyleGAN2-ADA to high-resolution medical imaging datasets. Our dataset is comprised of liver-containing axial slices from non-contrast and contrast-enhanced computed tomography (CT) scans. Additionally, we utilized four public datasets composed of various imaging modalities. We trained a StyleGAN2 network with transfer learning (from the Flickr-Faces-HQ dataset) and data augmentation (horizontal flipping and adaptive discriminator augmentation). The network’s generative quality was measured quantitatively with the Fréchet Inception Distance (FID) and qualitatively with a visual Turing test given to seven radiologists and radiation oncologists. The StyleGAN2-ADA network achieved a FID of 5.22 (±0.17) on our liver CT dataset. It also set new record FIDs of 10.78, 3.52, 21.17, and 5.39 on the publicly available SLIVER07, ChestX-ray14, ACDC, and Medical Segmentation Decathlon (brain tumors) datasets. In the visual Turing test, the clinicians rated generated images as real 42% of the time, approaching random guessing. Our computational ablation study revealed that transfer learning and data augmentation stabilize training and improve the perceptual quality of the generated images. We observed the FID to be consistent with human perceptual evaluation of medical images. Finally, our work found that StyleGAN2-ADA consistently produces high-quality results without hyperparameter searches or retraining.

Original languageEnglish (US)
Title of host publicationSimulation and Synthesis in Medical Imaging - 7th International Workshop, SASHIMI 2022, Held in Conjunction with MICCAI 2022, Proceedings
EditorsCan Zhao, David Svoboda, Jelmer M. Wolterink, Maria Escobar
PublisherSpringer Science and Business Media Deutschland GmbH
Pages142-153
Number of pages12
ISBN (Print)9783031169793
DOIs
StatePublished - 2022
Event7th International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore
Duration: Sep 18 2022Sep 18 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13570 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Country/TerritorySingapore
CitySingapore
Period9/18/229/18/22

Keywords

  • Data augmentation
  • Fréchet Inception Distance
  • StyleGAN2-ADA
  • Transfer learning
  • Visual turing test

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Evaluating the Performance of StyleGAN2-ADA on Medical Images'. Together they form a unique fingerprint.

Cite this