scholarly journals Semi-Automatic Segmentation of Vertebral Bodies in MR Images of Human Lumbar Spines

2018 ◽  
Vol 8 (9) ◽  
pp. 1586 ◽  
Author(s):  
Sewon Kim ◽  
Won Bae ◽  
Koichi Masuda ◽  
Christine Chung ◽  
Dosik Hwang
Keyword(s):  
Lumbar Spine ◽  
Vertebral Body ◽  
Automatic Segmentation ◽  
Dice Similarity Coefficient ◽  
Mr Images ◽  
Segmentation Accuracy ◽  
Body Segmentation ◽  
Spine Mr ◽  
Vertebral Bodies ◽  
Automatic Algorithm

We propose a semi-automatic algorithm for the segmentation of vertebral bodies in magnetic resonance (MR) images of the human lumbar spine. Quantitative analysis of spine MR images often necessitate segmentation of the image into specific regions representing anatomic structures of interest. Existing algorithms for vertebral body segmentation require heavy inputs from the user, which is a disadvantage. For example, the user needs to define individual regions of interest (ROIs) for each vertebral body, and specify parameters for the segmentation algorithm. To overcome these drawbacks, we developed a semi-automatic algorithm that considerably reduces the need for user inputs. First, we simplified the ROI placement procedure by reducing the requirement to only one ROI, which includes a vertebral body; subsequently, a correlation algorithm is used to identify the remaining vertebral bodies and to automatically detect the ROIs. Second, the detected ROIs are adjusted to facilitate the subsequent segmentation process. Third, the segmentation is performed via graph-based and line-based segmentation algorithms. We tested our algorithm on sagittal MR images of the lumbar spine and achieved a 90% dice similarity coefficient, when compared with manual segmentation. Our new semi-automatic method significantly reduces the user’s role while achieving good segmentation accuracy.

Vertebral Body Segmentation of Spine MR Images Using Superpixels

2015 ◽  
Author(s):  
Paulo Duarte Barbieri ◽  
Glauco Vitor Pedrosa ◽  
Agma Juci Machado Traina ◽  
Marcello Henrique Nogueira-Barbosa
Keyword(s):  
Vertebral Body ◽  
Mr Images ◽  
Body Segmentation ◽  
Spine Mr ◽  
Vertebral Body Segmentation

Semi-automatic segmentation of vertebral bodies in volumetric MR images using a statistical shape+pose model

2014 ◽  
Author(s):  
Amin Suzani ◽  
Abtin Rasoulian ◽  
Sidney Fels ◽  
Robert N. Rohling ◽  
Purang Abolmaesumi
Keyword(s):  
Automatic Segmentation ◽  
Mr Images ◽  
Statistical Shape ◽  
Vertebral Bodies

scholarly journals AN AUTOMATIC SEGMENTATION OF LIVER COLUME THAT CONTAINED DISCONNECTED REGION WITH DYNAMIC LOCAL NEIGHBOURHOOD REGION SIZES

2016 ◽  
Vol 3 (1) ◽  
pp. 129-148
Author(s):  
Puteri Suhaiza Sulaiman ◽  
Rahmita Wirza Rahmat ◽  
Ramlan Mahmod ◽  
Abdul Hamid Abdul Rashid
Keyword(s):  
Level Set ◽  
Automatic Segmentation ◽  
Dice Similarity Coefficient ◽  
Dynamic Localization ◽  
Hybrid Energy ◽  
Segmentation Accuracy ◽  
Average Factor ◽  
Liver Region ◽  
Level Set Framework ◽  
Localization Region

Segmentation of liver images containing disconnected regions has always been an overlooked problem. Previous works on liver segmentation either ignore this problem or use manual initialization when facing these disconnected regions. Therefore, in this paper we propose a liver level set (LLS) algorithm which is able to segment disconnected regions automatically. The LLS algorithm is based on level set framework together with hybrid energy minimization as the stopping function. By using the LLS algorithm in a looping manner, we allow the current liver boundary to inherit the topological changes from previous images in a 2.5D environment. We also conduct an experiment to obtain an average factor for dynamic localization region sizes based on liver anatomy to improve the segmentation accuracy. These dynamic localization region sizes ensure a more accurate segmentation when compared with manual segmentation. Our experiment gives a respective segmentation result with dice similarity coefficient (DSC) percentage of 87.5%. Plus, our LLS algorithm is able to segment all connected and disconnected liver region automatically and accurately.

scholarly journals DC2Anet: Generating Lumbar Spine MR Images from CT Scan Data Based on Semi-Supervised Learning

2019 ◽  
Vol 9 (12) ◽  
pp. 2521 ◽  
Author(s):  
Cheng-Bin Jin ◽  
Hakil Kim ◽  
Mingjie Liu ◽  
In Ho Han ◽  
Jae Il Lee ◽  
...  
Keyword(s):  
Lumbar Spine ◽  
Objective Function ◽  
Supervised Learning ◽  
Lumbar Disc ◽  
Mr Images ◽  
Disc Disease ◽  
Cardiac Pacemakers ◽  
Lumbar Disc Disease ◽  
Adversarial Network ◽  
Spine Mr

Magnetic resonance imaging (MRI) plays a significant role in the diagnosis of lumbar disc disease. However, the use of MRI is limited because of its high cost and significant operating and processing time. More importantly, MRI is contraindicated for some patients with claustrophobia or cardiac pacemakers due to the possibility of injury. In contrast, computed tomography (CT) scans are much less expensive, are faster, and do not face the same limitations. In this paper, we propose a method for estimating lumbar spine MR images based on CT images using a novel objective function and a dual cycle-consistent adversarial network (DC 2 Anet) with semi-supervised learning. The objective function includes six independent loss terms to balance quantitative and qualitative losses, enabling the generation of a realistic and accurate synthetic MR image. DC 2 Anet is also capable of semi-supervised learning, and the network is general enough for supervised or unsupervised setups. Experimental results prove that the method is accurate, being able to construct MR images that closely approximate reference MR images, while also outperforming four other state-of-the-art methods.

scholarly journals Retroperitoneal oblique corridor to the L2–S1 intervertebral discs: an MRI study

2016 ◽  
Vol 24 (2) ◽  
pp. 248-255 ◽  
Author(s):  
Diana M. Molinares ◽  
Timothy T. Davis ◽  
Daniel A. Fung
Keyword(s):  
Lumbar Spine ◽  
Intervertebral Disc ◽  
Vertebral Body ◽  
Psoas Muscle ◽  
Intervertebral Discs ◽  
Mr Images ◽  
Disc Space ◽  
Iliac Vessel ◽  
Medial Border ◽  
Transpsoas Approach

OBJECT The purpose of this study was to analyze MR images of the lumbar spine and document: 1) the oblique corridor at each lumbar disc level between the psoas muscle and the great vessels, and 2) oblique access to the L5–S1 disc space. Access to the lumbar spine without disruption of the psoas muscle could translate into decreased frequency of postoperative neurological complications observed after a transpsoas approach. The authors investigated the retroperitoneal oblique corridor of L2–S1 as a means of surgical access to the intervertebral discs. This oblique approach avoids the psoas muscle and is a safe and potentially superior alternative to the lateral transpsoas approach used by many surgeons. METHODS One hundred thirty-three MRI studies performed between May 4, 2012, and February 27, 2013, were randomly selected from the authors’ database. Thirty-three MR images were excluded due to technical issues or altered lumbar anatomy due to previous spine surgery. The oblique corridor was defined as the distance between the left lateral border of the aorta (or iliac artery) and the anterior medial border of the psoas. The L5–S1 oblique corridor was defined transversely from the midsagittal line of the inferior endplate of L-5 to the medial border of the left common iliac vessel (axial view) and vertically to the first vascular structure that crossed midline (sagittal view). RESULTS The oblique corridor measurements to the L2–5 discs have the following mean distances: L2–3 = 16.04 mm, L3–4 = 14.21 mm, and L4–5 = 10.28 mm. The L5–S1 corridor mean distance was 10 mm between midline and left common iliac vessel, and 10.13 mm from the first midline vessel to the inferior endplate of L-5. The bifurcation of the aorta and confluence of the vena cava were also analyzed in this study. The aortic bifurcation was found at the L-3 vertebral body in 2% of the MR images, at the L3–4 disc in 5%, at the L-4 vertebral body in 43%, at the L4–5 disc in 11%, and at the L-5 vertebral body in 9%. The confluence of the iliac veins was found at lower levels: 45% at the L-4 level, 19.39% at the L4–5 intervertebral disc, and 34% at the L-5 vertebral body. CONCLUSIONS An oblique corridor of access to the L2–5 discs was found in 90% of the MR images (99% access to L2–3, 100% access to L3–4, and 91% access to L4–5). Access to the L5–S1 disc was also established in 69% of the MR images analyzed. The lower the confluence of iliac veins, the less probable it was that access to the L5–S1 intervertebral disc space was observed. These findings support the use of lumbar MRI as a tool to predetermine the presence of an oblique corridor for access to the L2–S1 intervertebral disc spaces prior to lumbar spine surgery.

scholarly journals Deep learning-based segmentation of high-resolution computed tomography image data outperforms commonly used automatic bone segmentation methods

2021 ◽  
Author(s):  
Daniella M. Patton ◽  
Emilie N. Henning ◽  
Rob W. Goulet ◽  
Sean K. Carroll ◽  
Erin M.R. Bigelow ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Deep Learning ◽  
Femoral Neck ◽  
Vertebral Body ◽  
Automatic Segmentation ◽  
Image Data ◽  
Segmentation Methods ◽  
Unseen Data ◽  
Vertebral Bodies ◽  
Ct Data

Segmenting bone from background is required to quantify bone architecture in computed tomography (CT) image data. A deep learning approach using convolutional neural networks (CNN) is a promising alternative method for automatic segmentation. The study objectives were to evaluate the performance of CNNs in automatic segmentation of human vertebral body (micro-CT) and femoral neck (nano-CT) data and to investigate the performance of CNNs to segment data across scanners. Scans of human L1 vertebral bodies (microCT [North Star Imaging], n=28, 53μm3) and femoral necks (nano-CT [GE], n=28, 27μm3) were used for evaluation. Six slices were selected for each scan and then manually segmented to create ground truth masks (Dragonfly 4.0, ORS). Two-dimensional U-Net CNNs were trained in Dragonfly 4.0 with images of the [FN] femoral necks only, [VB] vertebral bodies only, and [F+V] combined CT data. Global (i.e., Otsu and Yen) and local (i.e., Otsu r = 100) thresholding methods were applied to each dataset. Segmentation performance was evaluated using the Dice index, a similarity metric of overlap. Kruskal-Wallis and Tukey-Kramer post-hoc tests were used to test for significant differences in the accuracy of segmentation methods. The FN U-Net had significantly higher Dice indices (i.e., better performance) than the global (Otsu: p=0.001; Yen: p=0.001) and local (Otsu [r=100]: p=0.001) thresholding methods and the VB U-Net (p=0.001) but there was no significant difference in model performance compared to the FN + VB U-net (p=0.783) on femoral neck image data. The VB U-net had significantly higher Dice coefficients than the global and local Otsu (p=0.001 for both) and FN U-Net (p=0.001) but not compared to the Yen (p=0.462) threshold or FN + VB U-net (p=0.783) on vertebral body image data. The results demonstrate that the U-net architecture outperforms common thresholding methods. Further, a network trained with bone data from a different system (i.e., different image acquisition parameters and voxel size) and a different anatomical site can perform well on unseen data. Finally, a network trained with combined datasets performed well on both datasets, indicating that a network can feasibly be trained with multiple datasets and perform well on varied image data.

scholarly journals Automatic Vertebral Body Segmentation using Semantic Segmentation

2019 ◽  
Vol 8 (4) ◽  
pp. 12163-12167
Keyword(s):  
Computed Tomography ◽  
Semantic Segmentation ◽  
Similarity Coefficient ◽  
Dice Similarity Coefficient ◽  
Challenging Problem ◽  
Average Accuracy ◽  
Body Segmentation ◽  
Vertebral Bodies ◽  
Segmentation Of Images ◽  
Main Components

Segmentation of vertebral bodies (VB) is a preliminary and useful step for the diagnosis of spine pathologies, deformations and fractures caused due to various reasons. We present a method to address this challenging problem of VB segmentation using a trending method – Semantic Segmentation (SS). The objective of semantic segmentation of images usually consisting of three main components - convolutions, downsampling, and upsampling layers is to mark every pixel of an image with a correlating class of what is being described. In this study, we developed a unique automatic semantic segmentation architecture to segment the VB from Computed Tomography (CT) images, and we compared our segmentation results with reference segmentations obtained by the experts. We evaluated the proposed method on a publicly available dataset and achieved an average accuracy of 94.16% and an average Dice Similarity Coefficient (DSC) of 93.51% for VB segmentation.

Sign in / Sign up

Export Citation Format

Share Document