Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring Foundation Models for Synthetic Medical Imaging: A Study on Chest X-Rays and Fine-Tuning Techniques

Davide Clode da Silva, Marina Musse Bernardes, Nathalia Giacomini Ceretta, Gabriel Vaz de Souza, Gabriel Fonseca Silva, Rafael Heitor Bordini, Soraia Raupp Musse

TL;DR

The potential of foundation models for generating realistic medical images, particularly chest x-rays, are explored, and how their performance improves with fine-tuning is assessed.

Abstract

Machine learning has significantly advanced healthcare by aiding in disease prevention and treatment identification. However, accessing patient data can be challenging due to privacy concerns and strict regulations. Generating synthetic, realistic data offers a potential solution for overcoming these limitations, and recent studies suggest that fine-tuning foundation models can produce such data effectively. In this study, we explore the potential of foundation models for generating realistic medical images, particularly chest x-rays, and assess how their performance improves with fine-tuning. We propose using a Latent Diffusion Model, starting with a pre-trained foundation model and refining it through various configurations. Additionally, we performed experiments with input from a medical professional to assess the realism of the images produced by each trained model.

Exploring Foundation Models for Synthetic Medical Imaging: A Study on Chest X-Rays and Fine-Tuning Techniques

TL;DR

The potential of foundation models for generating realistic medical images, particularly chest x-rays, are explored, and how their performance improves with fine-tuning is assessed.

Abstract

Machine learning has significantly advanced healthcare by aiding in disease prevention and treatment identification. However, accessing patient data can be challenging due to privacy concerns and strict regulations. Generating synthetic, realistic data offers a potential solution for overcoming these limitations, and recent studies suggest that fine-tuning foundation models can produce such data effectively. In this study, we explore the potential of foundation models for generating realistic medical images, particularly chest x-rays, and assess how their performance improves with fine-tuning. We propose using a Latent Diffusion Model, starting with a pre-trained foundation model and refining it through various configurations. Additionally, we performed experiments with input from a medical professional to assess the realism of the images produced by each trained model.
Paper Structure (7 sections, 3 figures, 2 tables)

This paper contains 7 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Proposed Model
  4. Dataset
  5. Fine-tuning
  6. Experimental Results
  7. Conclusion

Figures (3)

  • Figure 1: Sample images from the dataset and their respective annotations jaeger2014two. (a) presents a normal case, and (b) an abnormal case.
  • Figure 2: Set of normal chest x-ray images generated by the models presented in Table \ref{['tbl:configurations']}. All models except M1 were fine-tuned for 100 epochs on a dataset of 30 chest x-ray images. All images were generated using the prompt "healthy or normal human chest x-ray".
  • Figure 3: Set of abnormal chest x-ray images generated by the models presented in Table \ref{['tbl:configurations']}. All models except M0 were fine-tuned for 100 epochs on a dataset of 30 chest x-ray images. All images were generated using the prompt "Human chest x-ray with tuberculosis. Bilateral miliary nodules with Right Middle Lobe infiltrate. Right pleural effusion".