Deformable multi-modal medical image registration aligns the anatomical structures of different modalities to the same coordinate system through a spatial transformation. Due to the difficulties of collecting ground-truth registration labels, existing methods often adopt the unsupervised multi-modal image registration setting. However, it is hard to design satisfactory metrics to measure the similarity of multi-modal images, which heavily limits the multi-modal registration performance. Moreover, due to the contrast difference of the same organ in multi-modal images, it is difficult to extract and fuse the representations of different modal images. To address the above issues, we propose a novel unsupervised multi-modal adversarial registration framework that takes advantage of image-to-image translation to translate the medical image from one modality to another. In this way, we are able to use the well-defined uni-modal metrics to better train the models. Inside our framework, we propose two improvements to promote accurate registration. First, to avoid the translation network learning spatial deformation, we propose a geometry-consistent training scheme to encourage the translation network to learn the modality mapping solely. Second, we propose a novel semi-shared multi-scale registration network that extracts features of multi-modal images effectively and predicts multi-scale registration fields in an coarse-to-fine manner to accurately register the large deformation area. Extensive experiments on brain and pelvic datasets demonstrate the superiority of the proposed method over existing methods, revealing our framework has great potential in clinical application.

In this paper, we present a novel end-to-end neural model based on graph convolutional networks (GCN) for jointly extracting entities and relations between them. It divides the joint extraction into two sub-tasks, first detecting entity spans and identifying entity relations type simultaneously. To consider the complete interaction between entities and relations, we propose a novel relation-aware attention mechanism to obtain the relation representation between two entity spans. Therefore, a complete graph is constructed based on all extracted entity spans where the nodes are entity spans and the edges are relation representation. Besides, we improve original GCN to utilize both adjacent node features and edge information when encoding node feature. Experiments are conducted on two public datasets and our model outperforms all baseline methods.

By segmenting, modeling, and visualizing computed tomography sequence data, surgeons can better understand the spatial relationship of the intrahepatic vein, portal vein, artery, and tumor. Accurate preoperative understanding of liver anatomy and volume of the corresponding liver segment are necessary for anatomical liver resection. This paper gives an efficient and semiautomatic method for segmenting the liver in clinical cases. This method is based on Couinaud's theory and automatically divides the liver segment by portal vein branches. Considering the vascular variation of individual cases, the adjustment of the portal vein branches was provided on the basis of automatic segmentation, thus adaptability to various cases was implemented. For the final segmentation results, the portal vein blood supply of different liver segments can be confirmed by 3-D visualization, and the liver volume can be accurately estimated. Experiments show that this liver segmentation method has good clinical value.