Volumetric Conditioning Module to Control Pretrained Diffusion Models for 3D Medical Images

TL;DR

This paper introduces a lightweight volumetric conditioning module (VCM) for pretrained 3D diffusion models, enabling effective spatial control in medical image synthesis with limited data and computational resources.

Contribution

The paper proposes a novel VCM with an asymmetric U-Net architecture tailored for 3D medical images, facilitating conditional generation with less data and computation.

Findings

VCM effectively guides 3D medical image synthesis under various conditions.

It requires fewer training samples and less computational power.

Demonstrates potential for axial super-resolution in medical imaging.

Abstract

Spatial control methods using additional modules on pretrained diffusion models have gained attention for enabling conditional generation in natural images. These methods guide the generation process with new conditions while leveraging the capabilities of large models. They could be beneficial as training strategies in the context of 3D medical imaging, where training a diffusion model from scratch is challenging due to high computational costs and data scarcity. However, the potential application of spatial control methods with additional modules to 3D medical images has not yet been explored. In this paper, we present a tailored spatial control method for 3D medical images with a novel lightweight module, Volumetric Conditioning Module (VCM). Our VCM employs an asymmetric U-Net architecture to effectively encode complex information from various levels of 3D conditions, providing detailed guidance in image synthesis. To examine the applicability of spatial control methods and the effectiveness of VCM for 3D medical data, we conduct experiments under single- and multimodal conditions scenarios across a wide range of dataset sizes, from extremely small datasets with 10 samples to large datasets with 500 samples. The experimental results show that the VCM is effective for conditional generation and efficient in terms of requiring less training data and computational resources. We further investigate the potential applications for our spatial control method through axial super-resolution for medical images. Our code is available at \url{https://github.com/Ahn-Ssu/VCM}