Seyed-Parsa Hojjat1, Ismail Ayed2, Gregory J Garvin3, Kumaradevan Punithakumar4. 1. Deparment of Medical Imaging, Sunnybrook Health Sciences Centre, 2075 Bayview Ave., Toronto, ON, M4N 3M5, Canada. 2. Ecole de Technologie Superieure (ETS), 1100 Rue Notre-Dame O, Montreal, QC, H3C 1K3, Canada. ismail.benayed@etsmtl.ca. 3. Department of Medical Imaging, University of Western Ontario, 268 Grosvenor St, London, ON, N6A 4V2, Canada. 4. Department of Radiology and Diagnostic Imaging, University of Alberta, 3A5.039, 8440 - 112 St NW, Edmonton, AB, Canada.
Abstract
PURPOSE: This study investigates an efficient (nearly real-time) two-stage spine labeling algorithm that removes the need for an external training while being applicable to different types of MRI data and acquisition protocols. METHODS: Based solely on the image being labeled (i.e., we do not use training data), the first stage aims at detecting potential vertebra candidates following the optimization of a functional containing two terms: (i) a distribution-matching term that encodes contextual information about the vertebrae via a density model learned from a very simple user input, which amounts to a point (mouse click) on a predefined vertebra; and (ii) a regularization constraint, which penalizes isolated candidates in the solution. The second stage removes false positives and identifies all vertebrae and discs by optimizing a geometric constraint, which embeds generic anatomical information on the interconnections between neighboring structures. Based on generic knowledge, our geometric constraint does not require external training. RESULTS: We performed quantitative evaluations of the algorithm over a data set of 90 mid-sagittal MRI images of the lumbar spine acquired from 45 different subjects. To assess the flexibility of the algorithm, we used both T1- and T2-weighted images for each subject. A total of 990 structures were automatically detected/labeled and compared to ground-truth annotations by an expert. On the T2-weighted data, we obtained an accuracy of 91.6% for the vertebrae and 89.2% for the discs. On the T1-weighted data, we obtained an accuracy of 90.7% for the vertebrae and 88.1% for the discs. CONCLUSION: Our algorithm removes the need for external training while being applicable to different types of MRI data and acquisition protocols. Based on the current testing data, a subject-specific model density and generic anatomical information, our method can achieve competitive performances when applied to T1- and T2-weighted MRI images.
PURPOSE: This study investigates an efficient (nearly real-time) two-stage spine labeling algorithm that removes the need for an external training while being applicable to different types of MRI data and acquisition protocols. METHODS: Based solely on the image being labeled (i.e., we do not use training data), the first stage aims at detecting potential vertebra candidates following the optimization of a functional containing two terms: (i) a distribution-matching term that encodes contextual information about the vertebrae via a density model learned from a very simple user input, which amounts to a point (mouse click) on a predefined vertebra; and (ii) a regularization constraint, which penalizes isolated candidates in the solution. The second stage removes false positives and identifies all vertebrae and discs by optimizing a geometric constraint, which embeds generic anatomical information on the interconnections between neighboring structures. Based on generic knowledge, our geometric constraint does not require external training. RESULTS: We performed quantitative evaluations of the algorithm over a data set of 90 mid-sagittal MRI images of the lumbar spine acquired from 45 different subjects. To assess the flexibility of the algorithm, we used both T1- and T2-weighted images for each subject. A total of 990 structures were automatically detected/labeled and compared to ground-truth annotations by an expert. On the T2-weighted data, we obtained an accuracy of 91.6% for the vertebrae and 89.2% for the discs. On the T1-weighted data, we obtained an accuracy of 90.7% for the vertebrae and 88.1% for the discs. CONCLUSION: Our algorithm removes the need for external training while being applicable to different types of MRI data and acquisition protocols. Based on the current testing data, a subject-specific model density and generic anatomical information, our method can achieve competitive performances when applied to T1- and T2-weighted MRI images.
Entities:
Keywords:
Distribution matching; Geometric constraints; Magnetic resonance imaging (MRI); Regularization; Spine labelling
Authors: Yunliang Cai; Mark Landis; David T Laidley; Anat Kornecki; Andrea Lum; Shuo Li Journal: Comput Med Imaging Graph Date: 2016-04-08 Impact factor: 4.790
Authors: Zhijie Wang; Xiantong Zhen; KengYeow Tay; Said Osman; Walter Romano; Shuo Li Journal: IEEE Trans Med Imaging Date: 2014-10-29 Impact factor: 10.048
Authors: B Michael Kelm; Michael Wels; S Kevin Zhou; Sascha Seifert; Michael Suehling; Yefeng Zheng; Dorin Comaniciu Journal: Med Image Anal Date: 2012-12-01 Impact factor: 8.545
Authors: Brandon Miles; Ismail Ben Ayed; Max W K Law; Greg Garvin; Aaron Fenster; Shuo Li Journal: IEEE Trans Biomed Eng Date: 2013-01-29 Impact factor: 4.538