Fully automatic cervical vertebrae segmentation framework for X-ray images

Abstract

The cervical spine is a highly flexible anatomy and therefore vulnerable to injuries. Unfortunately, a large number of injuries in lateral cervical X-ray images remain undiagnosed due to human errors. Computer-aided injury detection has the potential to reduce the risk of misdiagnosis. Towards building an automatic injury detection system, in this paper, we propose a deep learning-based fully automatic framework for segmentation of cervical vertebrae in X-ray images. The framework first localizes the spinal region in the image using a deep fully convolutional neural network. Then vertebra centers are localized using a novel deep probabilistic spatial regression network. Finally, a novel shape-aware deep segmentation network is used to segment the vertebrae in the image. The framework can take an X-ray image and produce a vertebrae segmentation result without any manual intervention. Each block of the fully automatic framework has been trained on a set of 124 X-ray images and tested on another 172 images, all collected from real-life hospital emergency rooms. A Dice similarity coefficient of 0.84 and a shape error of 1.69 mm have been achieved.

Automatic Deep Feature Learning via Patch-Based Deep Belief Network for Vertebrae Segmentation in CT Images

“…Y. Li et al [15] 81.0 S. F. Qadri et al [44] 85.0 A. Seitel et al [57] 83.0 A. Sekuboyina et al [58] 87.0 S. M. M. R. Al Arif et al [41] 84.0 Our method 86.1…”

“…However, this model fails in the structural consistency of the vertebrae parts segmentation or sometimes entire vertebra from starting and ending slices. Another deep learning method [41] is proposed for cervical vertebrae segmentation from X-ray images and achieved DICE similarity coefficient of 84%; however, the patch-based center localization method has the limitation of not knowing which center belongs to which vertebra. Our method is implemented in MATLAB and thus consumes time to segment the vertebrae.…”

“…In a previous study [40], the corner detection of cervical vertebrae using the patch-based methodology in X-ray images is proposed. This work is extended for cervical vertebra segmentation in X-ray images by using a deep fully convolutional neural network (CNN) and has achieved a DICE similarity coefficient of 84% with a shape error of 1.69 mm [41]. In this work, twelve patches around the vertebra are generated to provide the spatial context, and the vertebra corners are detected using a machine learning method.…”

Artificial Intelligence for Automatic Measurement of Sagittal Vertical Axis Using ResUNet Framework

“…There were also studies on radiographs: for biplanar radiographs, including A-P and lateral view, Gallbusera et al [18] used a database collected using the EOS™ imaging system [19] and trained CNN models for each of the landmarks. For lateral spine radiographs, Al Arif et al [20] applied a UNet model for the localization of cervical vertebral centers.…”

Automated Measurement of Foot Deformities: Flatfoot, High Arch, Calcaneal Fracture

“…Despite the common belief that x-ray is being replaced by more modern diagnostic techniques such as computed tomography (CT) and magnetic resonance imaging (MRI), it still holds a vital role in the art of medical diagnostic (Arif et al, 2018;Wojciechowski et al, 2016). As is stated in Arslan et al (2014), x-ray radiographs are fast, easily accessible, and expose patients to less radiation than CT.…”