Anatomy-Preserving Latent Diffusion for Generation of Brain Segmentation Masks with Ischemic Infarct

TL;DR

This paper introduces an anatomy-preserving generative framework using a latent diffusion model and VAE for synthesizing realistic brain segmentation masks, including ischemic infarcts, to address data scarcity in medical imaging.

Contribution

It presents a novel combination of a VAE and diffusion model in latent space for anatomy-aware mask generation, improving realism and structural integrity over pixel-space methods.

Findings

Generated masks preserve brain anatomy and tissue semantics.

The method avoids structural artifacts common in pixel-space models.

Synthetic masks show realistic variability and lesion control.

Abstract

The scarcity of high-quality segmentation masks remains a major bottleneck for medical image analysis, particularly in non-contrast CT (NCCT) neuroimaging, where manual annotation is costly and variable. To address this limitation, we propose an anatomy-preserving generative framework for the unconditional synthesis of multi-class brain segmentation masks, including ischemic infarcts. The proposed approach combines a variational autoencoder trained exclusively on segmentation masks to learn an anatomical latent representation, with a diffusion model operating in this latent space to generate new samples from pure noise. At inference, synthetic masks are obtained by decoding denoised latent vectors through the frozen VAE decoder, with optional coarse control over lesion presence via a binary prompt. Qualitative results show that the generated masks preserve global brain anatomy, discrete tissue semantics, and realistic variability, while avoiding the structural artifacts commonly observed in pixel-space generative models. Overall, the proposed framework offers a simple and scalable solution for anatomy-aware mask generation in data-scarce medical imaging scenarios.