Domain-Adaptive 3D Medical Image Synthesis: An Efficient Unsupervised Approach
This work addresses the limited applicability of synthesis models in clinical routines due to domain shifts, representing an incremental advance in domain adaptation for medical imaging.
The paper tackles the problem of domain shift in 3D medical image synthesis by proposing an unsupervised domain adaptation approach using a 2D variational autoencoder to approximate 3D distributions, resulting in significantly improved synthesis accuracy on unseen domains.
Medical image synthesis has attracted increasing attention because it could generate missing image data, improving diagnosis and benefits many downstream tasks. However, so far the developed synthesis model is not adaptive to unseen data distribution that presents domain shift, limiting its applicability in clinical routine. This work focuses on exploring domain adaptation (DA) of 3D image-to-image synthesis models. First, we highlight the technical difference in DA between classification, segmentation and synthesis models. Second, we present a novel efficient adaptation approach based on 2D variational autoencoder which approximates 3D distributions. Third, we present empirical studies on the effect of the amount of adaptation data and the key hyper-parameters. Our results show that the proposed approach can significantly improve the synthesis accuracy on unseen domains in a 3D setting. The code is publicly available at https://github.com/WinstonHuTiger/2D_VAE_UDA_for_3D_sythesis