Automatic Localization of the Lumbar Vertebral Landmarks in CT Images with Context Features
File(s)
Accepted version
Author(s)
Damopoulos, D
Glocker, B
Zheng, G
Type
Conference Paper
Abstract
A recent research direction for the localization of anatomical landmarks
with learning-based methods is to explore ways to enrich the trained models
with context information. Lately, the addition of context features in regression-based
approaches has been tried in the literature. In this work, a method is
presented for the addition of context features in a regression setting where the
locations of many vertebral landmarks are regressed all at once. As this method
relies on the knowledge of the centers of the vertebral bodies (VBs), an automatic,
endplate-based approach for the localization of the VB centers is also presented.
The proposed methods are evaluated on a dataset of 28 lumbar-focused CT images.
The VB localization method detects all of the lumbar VBs of the testing set
with a mean localization error of 3.2 mm. The multi-landmark localization
method is tested on the task of localizing the tips of all the inferior articular processes
of the lumbar vertebrae, in addition to their VB centers. The proposed
method detects these landmarks with a mean localization error of 3.0 mm.
with learning-based methods is to explore ways to enrich the trained models
with context information. Lately, the addition of context features in regression-based
approaches has been tried in the literature. In this work, a method is
presented for the addition of context features in a regression setting where the
locations of many vertebral landmarks are regressed all at once. As this method
relies on the knowledge of the centers of the vertebral bodies (VBs), an automatic,
endplate-based approach for the localization of the VB centers is also presented.
The proposed methods are evaluated on a dataset of 28 lumbar-focused CT images.
The VB localization method detects all of the lumbar VBs of the testing set
with a mean localization error of 3.2 mm. The multi-landmark localization
method is tested on the task of localizing the tips of all the inferior articular processes
of the lumbar vertebrae, in addition to their VB centers. The proposed
method detects these landmarks with a mean localization error of 3.0 mm.
Date Issued
2017-09-10
Date Acceptance
2017-07-11
Citation
2017
Copyright Statement
© 2017 5th MICCAI Workshop on Computational Methods and Clinical
Applications in Musculoskeletal Imaging – MICCAI-MSKI2017. All articles published herein are protected by copyright, which covers the exclusive rights to
reproduce and distribute the articles, as well all translation rights. No materials published herein
may be reproduced or stored electronically without first obtaining written permission from the
authors
Applications in Musculoskeletal Imaging – MICCAI-MSKI2017. All articles published herein are protected by copyright, which covers the exclusive rights to
reproduce and distribute the articles, as well all translation rights. No materials published herein
may be reproduced or stored electronically without first obtaining written permission from the
authors
Source
Computational Methods and Clinical Applications in Musculoskeletal Imaging (MSKI)
Publication Status
Published
Start Date
2017-09-10
Finish Date
2017-09-10
Coverage Spatial
Quebec City, Canada