Brain Imaging Generation with Latent Diffusion Models
read the original abstract
Deep neural networks have brought remarkable breakthroughs in medical image analysis. However, due to their data-hungry nature, the modest dataset sizes in medical imaging projects might be hindering their full potential. Generating synthetic data provides a promising alternative, allowing to complement training datasets and conducting medical image research at a larger scale. Diffusion models recently have caught the attention of the computer vision community by producing photorealistic synthetic images. In this study, we explore using Latent Diffusion Models to generate synthetic images from high-resolution 3D brain images. We used T1w MRI images from the UK Biobank dataset (N=31,740) to train our models to learn about the probabilistic distribution of brain images, conditioned on covariables, such as age, sex, and brain structure volumes. We found that our models created realistic data, and we could use the conditioning variables to control the data generation effectively. Besides that, we created a synthetic dataset with 100,000 brain images and made it openly available to the scientific community.
This paper has not been read by Pith yet.
Forward citations
Cited by 4 Pith papers
-
BrainG3N: A Dual-Purpose Tokenizer for Controllable 3D Brain MRI Generation
A volumetric MAE tokenizer decouples clinical embedding from reconstruction to support both 23-task linear probing and conditional 3D brain MRI generation via DiT.
-
Do Synthetic Brain MRIs Reliably Improve Tumour Classification? A StyleGAN2-ADA Class-Plane Augmentation Study on BRISC 2025
Filtered 1:1 synthetic MRI augmentation improved MobileViTV2 tumour classification accuracy by 1.02% on BRISC 2025 while providing no benefit to random forest and non-significant gains for CNN.
-
Triple-Phase Sequential Fusion Network for Hepatobiliary Phase Liver MRI Synthesis
TriPF-Net synthesizes hepatobiliary phase MRI images from pre-HBP sequences using adaptive triple-phase fusion and clinical variables, reporting MAE 10.65-12.41, PSNR 23.11-23.27, and SSIM 0.76-0.78 on internal and ex...
-
DustNET: enabling machine learning and AI models of dusty plasmas
DustNET is proposed as a shared dataset to train machine learning models that complement traditional physics equations for predictive modeling of dusty plasmas across laboratory and natural scales.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.