Multi-Conditioned Denoising Diffusion Probabilistic Model (mDDPM) for Medical Image Synthesis

TL;DR

This paper introduces a multi-conditioned denoising diffusion probabilistic model (mDDPM) for generating realistic, annotated lung CT images with controlled specifications, surpassing existing models in anatomical accuracy.

Contribution

The paper presents a novel multi-conditioned DDPM framework that enables controlled, high-fidelity medical image synthesis with annotations, improving anatomical consistency over prior models.

Findings

Generated images convincingly fool experts as real.

Controlled generation surpasses state-of-the-art in anatomical accuracy.

Model trained on large datasets demonstrates superior realism.

Abstract

Medical imaging applications are highly specialized in terms of human anatomy, pathology, and imaging domains. Therefore, annotated training datasets for training deep learning applications in medical imaging not only need to be highly accurate but also diverse and large enough to encompass almost all plausible examples with respect to those specifications. We argue that achieving this goal can be facilitated through a controlled generation framework for synthetic images with annotations, requiring multiple conditional specifications as input to provide control. We employ a Denoising Diffusion Probabilistic Model (DDPM) to train a large-scale generative model in the lung CT domain and expand upon a classifier-free sampling strategy to showcase one such generation framework. We show that our approach can produce annotated lung CT images that can faithfully represent anatomy, convincingly fooling experts into perceiving them as real. Our experiments demonstrate that controlled generative frameworks of this nature can surpass nearly every state-of-the-art image generative model in achieving anatomical consistency in generated medical images when trained on comparable large medical datasets.