eduzhai > Applied Sciences > Engineering >

Self-Supervised Ultrasound to MRI Fetal Brain Image Synthesis

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: Fetal brain magnetic resonance imaging (MRI) offers exquisite images of thedeveloping brain but is not suitable for second-trimester anomaly screening,for which ultrasound (US) is employed. Although expert sonographers are adeptat reading US images, MR images which closely resemble anatomical images aremuch easier for non-experts to interpret. Thus in this paper we propose togenerate MR-like images directly from clinical US images. In medical imageanalysis such a capability is potentially useful as well, for instance forautomatic US-MRI registration and fusion. The proposed model is end-to-endtrainable and self-supervised without any external annotations. Specifically,based on an assumption that the US and MRI data share a similar anatomicallatent space, we first utilise a network to extract the shared latent features,which are then used for MRI synthesis. Since paired data is unavailable for ourstudy (and rare in practice), pixel-level constraints are infeasible to apply.We instead propose to enforce the distributions to be statisticallyindistinguishable, by adversarial learning in both the image domain and featurespace. To regularise the anatomical structures between US and MRI duringsynthesis, we further propose an adversarial structural constraint. A newcross-modal attention technique is proposed to utilise non-local spatialinformation, by encouraging multi-modal knowledge fusion and propagation. Weextend the approach to consider the case where 3D auxiliary information (e.g.,3D neighbours and a 3D location index) from volumetric data is also available,and show that this improves image synthesis. The proposed approach is evaluatedquantitatively and qualitatively with comparison to real fetal MR images andother approaches to synthesis, demonstrating its feasibility of synthesisingrealistic MR images.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...