Automatic Localization and Identification of Vertebrae in Spine CT via a Joint Learning Model with Deep Neural Networks
Accurate localization and identification of vertebrae in 3D spinal images is essential for many clinical tasks. However, automatic localization and identification of vertebrae remains challenging due to similar appearance of vertebrae, abnormal pathologic
- PDF / 1,364,972 Bytes
- 8 Pages / 439.363 x 666.131 pts Page_size
- 64 Downloads / 182 Views
Dept. of Computer Science and Engineering, The Chinese University of Hong Kong 2 College of Computer Science and Technology, Zhejiang University, China 3 School of Medicine, Shenzhen University, China 4 Prince of Wales Hospital, The Chinese University of Hong Kong
Abstract. Accurate localization and identification of vertebrae in 3D spinal images is essential for many clinical tasks. However, automatic localization and identification of vertebrae remains challenging due to similar appearance of vertebrae, abnormal pathological curvatures and image artifacts induced by surgical implants. Traditional methods relying on hand-crafted low level features and/or a priori knowledge usually fail to overcome these challenges on arbitrary CT scans. We present a robust and efficient approach to automatically locating and identifying vertebrae in 3D CT volumes by exploiting high level feature representations with deep convolutional neural network (CNN). A novel joint learning model with CNN (J-CNN) is proposed by considering both the appearance of vertebrae and the pairwise conditional dependency of neighboring vertebrae. The J-CNN can effectively identify the type of vertebra and eliminate false detections based on a set of coarse vertebral centroids generated from a random forest classifier. Furthermore, the predicted centroids are refined by a shape regression model. Our approach was quantitatively evaluated on the dataset of MICCAI 2014 Computational Challenge on Vertebrae Localization and Identification. Compared with the state-of-the-art method [1], our approach achieved a large margin with 10.12% improvement of the identification rate and smaller localization errors.
1
Introduction
Vertebrae serve as the essential anatomical landmarks and provide an important prior structure of spinal shape. Accurate localization and identification of vertebrae in 3D images such as CT and MRI is significant for many subsequent tasks including pathological diagnosis, surgical planning and post-operative assessment. Moreover, it can benefit many other applications including vertebral body segmentation [2], idiopathic scoliosis diagnosis [3], and intervertebral disc labelling [4], to name a few.
Joint first authors.
c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part I, LNCS 9349, pp. 515–522, 2015. DOI: 10.1007/978-3-319-24553-9_63
516
H. Chen et al.
(a)
(b)
(c)
(d)
Fig. 1. Challenges of vertebrae localization and identification in 3D CT volume (vertebral centroids are marked as red crosses): (a) similar appearance of vertebrae; (b) image artifacts induced by the implanted pedicle screws; (c)(d) pathological cases of scoliosis.
In clinical practice, manual labelling of vertebrae is time consuming and subjective with limited reproducibility [5]. Therefore, an accurate and reliable automatic localization and identification method can greatly improve the examination efficiency and reliability. However, developing such a method faces several challenges. First, vertebrae often carry similar morpholog
Data Loading...
