Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Medical Imaging Complexity and its Effects on GAN Performance

William Cagas, Chan Ko, Blake Hsiao, Shryuk Grandhi, Rishi Bhattacharya, Kevin Zhu, Michael Lam

TL;DR

This work experimentally establishes benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities, and analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy in information theory.

Abstract

The proliferation of machine learning models in diverse clinical applications has led to a growing need for high-fidelity, medical image training data. Such data is often scarce due to cost constraints and privacy concerns. Alleviating this burden, medical image synthesis via generative adversarial networks (GANs) emerged as a powerful method for synthetically generating photo-realistic images based on existing sets of real medical images. However, the exact image set size required to efficiently train such a GAN is unclear. In this work, we experimentally establish benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities. We analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy in information theory. For our pipeline, we conduct experiments with two state-of-the-art GANs, StyleGAN 3 and SPADE-GAN, trained on multiple medical imaging datasets with variable sample sizes. Across both GANs, general performance improved with increasing training set size but suffered with increasing complexity.

Medical Imaging Complexity and its Effects on GAN Performance

TL;DR

This work experimentally establishes benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities, and analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy in information theory.

Abstract

The proliferation of machine learning models in diverse clinical applications has led to a growing need for high-fidelity, medical image training data. Such data is often scarce due to cost constraints and privacy concerns. Alleviating this burden, medical image synthesis via generative adversarial networks (GANs) emerged as a powerful method for synthetically generating photo-realistic images based on existing sets of real medical images. However, the exact image set size required to efficiently train such a GAN is unclear. In this work, we experimentally establish benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities. We analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy in information theory. For our pipeline, we conduct experiments with two state-of-the-art GANs, StyleGAN 3 and SPADE-GAN, trained on multiple medical imaging datasets with variable sample sizes. Across both GANs, general performance improved with increasing training set size but suffered with increasing complexity.

Paper Structure

This paper contains 17 sections, 5 equations, 3 figures.

Table of Contents

  1. Introduction
  2. Background
  3. Generative Adversarial Networks (GANs)
  4. Image Complexity
  5. Methodology
  6. Image Complexity Metric
  7. GAN Selection
  8. GAN Pipeline
  9. Preprocessing
  10. Training and Generation
  11. Evaluation
  12. Experimental Results
  13. Datasets
  14. Discussion
  15. Broader Impacts
  16. ...and 2 more sections

Figures (3)

  • Figure 1: Comparison between original images and synthetic images from StyleGAN 3 and SPADE-GAN based on variable image set sizes.
  • Figure 2: Delentropy distributions across each medical image dataset. A higher mean delentropy $\mu$ indicates a dataset with more complex images.
  • Figure 3: Fréchet Inception Distance (FID) curves comparing StyleGAN 3 and SPADE-GAN across each medical image dataset with varying sample sizes. Lower FID scores correspond to higher fidelity synthetic images.