Echo from noise: synthetic ultrasound image generation using diffusion models for real image segmentation
This addresses data scarcity in medical imaging for researchers and clinicians, though it is an incremental application of diffusion models to a specific domain.
The authors tackled the problem of limited real ultrasound data for training segmentation models by generating synthetic ultrasound images using diffusion models guided by semantic labels. Their approach achieved mean Dice scores of 88.6%, 91.9%, and 85.2% for cardiac structures, representing relative improvements of 9.2%, 3.3%, and 13.9% over previous state-of-the-art methods.
We propose a novel pipeline for the generation of synthetic ultrasound images via Denoising Diffusion Probabilistic Models (DDPMs) guided by cardiac semantic label maps. We show that these synthetic images can serve as a viable substitute for real data in the training of deep-learning models for ultrasound image analysis tasks such as cardiac segmentation. To demonstrate the effectiveness of this approach, we generated synthetic 2D echocardiograms and trained a neural network for segmenting the left ventricle and left atrium. The performance of the network trained on exclusively synthetic images was evaluated on an unseen dataset of real images and yielded mean Dice scores of 88.6 $\pm 4.91$ , 91.9 $\pm 4.22$, 85.2 $\pm 4.83$ \% for left ventricular endocardium, epicardium and left atrial segmentation respectively. This represents a relative increase of $9.2$, $3.3$ and $13.9$ \% in Dice scores compared to the previous state-of-the-art. The proposed pipeline has potential for application to a wide range of other tasks across various medical imaging modalities.