Deep Generative Models for 3D Medical Image Synthesis

TL;DR

This paper reviews deep generative models like VAEs, GANs, and DDMs for synthesizing 3D medical images, highlighting their principles, applications, and challenges in clinical contexts.

Contribution

It provides a comprehensive overview of recent advances, applications, and evaluation metrics for deep generative models in 3D medical image synthesis.

Findings

GANs and VAEs enable realistic image generation

Diffusion models show promise in image quality

Evaluation metrics assess fidelity, diversity, and privacy

Abstract

Deep generative modeling has emerged as a powerful tool for synthesizing realistic medical images, driving advances in medical image analysis, disease diagnosis, and treatment planning. This chapter explores various deep generative models for 3D medical image synthesis, with a focus on Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Denoising Diffusion Models (DDMs). We discuss the fundamental principles, recent advances, as well as strengths and weaknesses of these models and examine their applications in clinically relevant problems, including unconditional and conditional generation tasks like image-to-image translation and image reconstruction. We additionally review commonly used evaluation metrics for assessing image fidelity, diversity, utility, and privacy and provide an overview of current challenges in the field.