Estimation of 3D rotation of femur in 2D hip radiographs
(2012) In Journal of Biomechanics 45(13). p.2279-2283- Abstract
- Femoral radiographs are affected by the degree of rotation of the femur with respect to the plane of projection. We aimed to determine the 3D rotation of the proximal femur in 2D radiographs. A 3D Statistical Appearance Model (SAM), which was built from CT images of cadaver proximal femurs (n = 33) was randomly sampled to form a training set of 500 bones. Nineteen clinical CT images were collected for testing. All CT images were rotated to +/- 20 degrees in 2 degrees division around the shaft axis, +/- 10 degrees around medial-lateral axis, and by simultaneous rotation of both axes (+/- 16 degrees and +/- 8 degrees around shaft and medial-lateral axes). In each orientation, a 2D projection was recorded for generating a 2D SAM. The outcome... (More)
- Femoral radiographs are affected by the degree of rotation of the femur with respect to the plane of projection. We aimed to determine the 3D rotation of the proximal femur in 2D radiographs. A 3D Statistical Appearance Model (SAM), which was built from CT images of cadaver proximal femurs (n = 33) was randomly sampled to form a training set of 500 bones. Nineteen clinical CT images were collected for testing. All CT images were rotated to +/- 20 degrees in 2 degrees division around the shaft axis, +/- 10 degrees around medial-lateral axis, and by simultaneous rotation of both axes (+/- 16 degrees and +/- 8 degrees around shaft and medial-lateral axes). In each orientation, a 2D projection was recorded for generating a 2D SAM. The outcome parameters of the 2D SAM were used as input for a linear regression model and an artificial neural network to predict the rotation. The artificial neural network estimated the rotation more accurately than the linear regression. For artificial neural networks the mean errors were 4.0 degrees and 2.0 degrees around the shaft and medial-lateral axes, respectively. For an individual radiograph, the confidence interval of estimation was still relatively large. However, this method has high potential to differentiate the amount of rotations in two image sets. (C) 2012 Elsevier Ltd. All rights reserved. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/3191440
- author
- Vaananen, Sami P. ; Isaksson, Hanna LU ; Waarsing, Jan H. ; Zadpoor, Amir Abbas ; Jurvelin, Jukka S. and Weinans, Harrie
- organization
- publishing date
- 2012
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Rotation of femur, X-ray, Statistical appearance model, Artificial, neural networks
- in
- Journal of Biomechanics
- volume
- 45
- issue
- 13
- pages
- 2279 - 2283
- publisher
- Elsevier
- external identifiers
-
- wos:000308854400014
- scopus:84865183768
- pmid:22796003
- ISSN
- 1873-2380
- DOI
- 10.1016/j.jbiomech.2012.06.007
- language
- English
- LU publication?
- yes
- id
- d97ba5ee-4588-44b1-91c8-8d3c5240cfcd (old id 3191440)
- date added to LUP
- 2016-04-01 10:16:31
- date last changed
- 2023-08-30 22:22:46
@article{d97ba5ee-4588-44b1-91c8-8d3c5240cfcd, abstract = {{Femoral radiographs are affected by the degree of rotation of the femur with respect to the plane of projection. We aimed to determine the 3D rotation of the proximal femur in 2D radiographs. A 3D Statistical Appearance Model (SAM), which was built from CT images of cadaver proximal femurs (n = 33) was randomly sampled to form a training set of 500 bones. Nineteen clinical CT images were collected for testing. All CT images were rotated to +/- 20 degrees in 2 degrees division around the shaft axis, +/- 10 degrees around medial-lateral axis, and by simultaneous rotation of both axes (+/- 16 degrees and +/- 8 degrees around shaft and medial-lateral axes). In each orientation, a 2D projection was recorded for generating a 2D SAM. The outcome parameters of the 2D SAM were used as input for a linear regression model and an artificial neural network to predict the rotation. The artificial neural network estimated the rotation more accurately than the linear regression. For artificial neural networks the mean errors were 4.0 degrees and 2.0 degrees around the shaft and medial-lateral axes, respectively. For an individual radiograph, the confidence interval of estimation was still relatively large. However, this method has high potential to differentiate the amount of rotations in two image sets. (C) 2012 Elsevier Ltd. All rights reserved.}}, author = {{Vaananen, Sami P. and Isaksson, Hanna and Waarsing, Jan H. and Zadpoor, Amir Abbas and Jurvelin, Jukka S. and Weinans, Harrie}}, issn = {{1873-2380}}, keywords = {{Rotation of femur; X-ray; Statistical appearance model; Artificial; neural networks}}, language = {{eng}}, number = {{13}}, pages = {{2279--2283}}, publisher = {{Elsevier}}, series = {{Journal of Biomechanics}}, title = {{Estimation of 3D rotation of femur in 2D hip radiographs}}, url = {{http://dx.doi.org/10.1016/j.jbiomech.2012.06.007}}, doi = {{10.1016/j.jbiomech.2012.06.007}}, volume = {{45}}, year = {{2012}}, }