Segmentation-Guided Knee Radiograph Generation using Conditional Diffusion Models

TL;DR

This paper introduces two strategies using conditional diffusion models to generate realistic knee radiographs from segmentations, aiding medical dataset augmentation with improved accuracy over traditional methods.

Contribution

The paper presents novel conditional diffusion model techniques for synthesizing knee radiographs from segmentation data, outperforming existing approaches.

Findings

Conditional training yields more realistic images than conditional sampling.

Both methods produce images consistent with input segmentations.

Proposed models outperform conventional U-Net in quality.

Abstract

Deep learning-based medical image processing algorithms require representative data during development. In particular, surgical data might be difficult to obtain, and high-quality public datasets are limited. To overcome this limitation and augment datasets, a widely adopted solution is the generation of synthetic images. In this work, we employ conditional diffusion models to generate knee radiographs from contour and bone segmentations. Remarkably, two distinct strategies are presented by incorporating the segmentation as a condition into the sampling and training process, namely, conditional sampling and conditional training. The results demonstrate that both methods can generate realistic images while adhering to the conditioning segmentation. The conditional training method outperforms the conditional sampling method and the conventional U-Net.