As part of the VerSe 2019 challenge a large dataset was provided addressing the previous severe shortage of publicly-available, large, accurately annotated CT spine data in the community by releasing 160 CT image series and their voxel-level annotations comprised of a large variety in fields of view and spatial resolutions as well as spinal and vertebral pathologies 20.īuilding on the data, experience, and learning from the VerSe 2019 challenge, we proposed to organise a second iteration of the vertebrae segmentation challenge at the MICCAI 2020 in Lima, Peru. The segmentation challenge received considerable participation from the scientific community with more than 250 registrations and 20 participating teams 19. Aiming at the task of improving automated quantification of spinal morphology and pathology by vertebral labelling and segmentation, the first iteration of the “Large Scale Vertebrae Segmentation Challenge” (VerSe 2019) 17, 19 was held at the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2019, Shenzhen, China). However, most of these approaches are largely data dependent, as the algorithms require extensive datasets with corresponding metadata for the development, training and validation to enable efficient models 18. Various applications of computer algorithms have shown great potential to detect vertebral fractures and to measure bone mineral density (BMD) 13, 14, 15, 16, 17. In spinal imaging, different deep learning approaches have been used for vertebral labelling and segmentation tasks in the form of convolutional neural networks (CNN), graph convolutional networks (GCN) or point clouds (PC) to analyse bone structures 8, 9, 10, 11, 12, 13. Numerous applications of computer-aided diagnostics (CADx) are currently being developed beginning to gradually reshape the future of radiological clinical practice and research 1, 2, 3, 4, 5, 6, 7. Metadata includes vertebral labelling information, voxel-level segmentation masks obtained with a human-machine hybrid algorithm and anatomical ratings, to enable the development and benchmarking of robust and accurate segmentation algorithms. VerSe 2020 comprises annotated spine computed tomography (CT) images from 300 subjects with 4142 fully visualized and annotated vertebrae, collected across multiple centres from four different scanner manufacturers, enriched with cases that exhibit anatomical variants such as enumeration abnormalities (n = 77) and transitional vertebrae (n = 161). Building on that experience, we report on the largely increased VerSe 2020 dataset and results from the second iteration of the VerSe challenge (MICCAI 2020, Lima, Peru). The first “Large Scale Vertebrae Segmentation Challenge” (VerSe 2019) showed that these perform well on normal anatomy, but fail in variants not frequently present in the training dataset. In spine imaging, several atlas- and shape-based as well as deep learning segmentation algorithms have been proposed, allowing for subsequent automated analysis of morphology and pathology. ![]() With the advent of deep learning algorithms, fully automated radiological image analysis is within reach.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |