Accurate vertebral segmentation is important for image-related assessment of spine pathologies. Recent spine segmentation techniques focus on deep learning-based methods. However, they typically require a large training dataset with accurate annotations, which may not be feasible in practice. To mitigate the challenge, here we propose a data augmentation scheme based on a single spine to incorporate biomechanically constrained intervertebral motion (rotation and translation) to generate a large training dataset (N=3000). Vertebrae are then segmented using a point cloud-based deep learning architecture, PointNet++, which appears not to have been applied to spine segmentation before. Spine testing samples (N=90) are first generated by applying our data augmentation technique to three separate spines. Segmentation performances are compared with two baseline data augmentation methods that do not include intervertebral motion. Then, we further evaluate our technique based on 8 spine samples from actual image acquisitions. In both cases, we show that incorporation of vertebral level-wise data augmentation improves segmentation accuracy in terms of mean Dice coefficient (e.g., 0.932 vs. 0.902 for augmented testing samples and 0.940 vs. 0.924 for acquired testing samples). These results suggest that our vertebral level-wise data augmentation is useful to facilitate deep learning-based spine segmentation, which is important especially when it is not feasible to generate accurate annotations for a large number of training samples.
In open spine surgery, the accuracy of image guidance is compromised by alignment change between supine preoperative CT images (pCT) and prone intraoperative positioning. We have developed a level-wise registration framework to compensate for the intervertebral motion by updating pCT to match with intraoperative stereovision (iSV) data of the exposed spine. In this study, we compared performance of the iSV image updating system in different lengths of exposure using retrospective data from one cadaver pig specimen. Specifically, L1 to L6 were exposed and 3 metallic mini-screws were implanted on each level as “ground truth” locations. The spine was positioned supine to acquire pCT, and then positioned prone to acquire iSV using a hand-held iSV device. One image pair of iSV was acquired from each exposed vertebra. Three exposure lengths were evaluated by selecting data from corresponding levels to compare performance: 6 levels, 4 levels, and 3 levels. Accuracy of iSV updating was assessed through point-to-point registration error (ppRE) using mini-screw locations, and the average accuracy was 1.26±0.77 mm, 1.54±0.62 mm, and 1.38±0.44 mm, for the three exposure lengths, respectively. The time cost was ~10-15 min and similar in all three exposure sizes. Results indicate that performance of iSV image updating was similar in different lengths of exposure, and the accuracy was within clinically acceptable range (2 mm).
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.