Vertebral classification using localized pathology-related shape model

R. Zewail; A. Elsafi; N. Durdle

doi:10.1117/12.770314

26 March 2008 Vertebral classification using localized pathology-related shape model

R. Zewail, A. Elsafi, N. Durdle

Proceedings Volume 6914, Medical Imaging 2008: Image Processing; 69144P (2008) https://doi.org/10.1117/12.770314
Event: Medical Imaging, 2008, San Diego, California, United States

Abstract

Radiographs of the spine are frequently examined for assessment of vertebral abnormalities. Features like osteophytes (bony growth of vertebra's corners), and disc space narrowing are often used as visual evidence of osteoarthris or degenerative joint disease. These symptoms result in remarkable changes in the shapes of the vertebral body. Statistical analysis of anatomical structure has recently gained increased popularity within the medical imaging community, since they have the potential to enhance the automated diagnosis process. In this paper, we present a novel method for computer-assisted vertebral classification using a localized, pathology-related shape model. The new classification scheme is able to assess the condition of multiple vertebrae simultaneously, hence is possible to directly classify the whole spine anatomy according to the condition of interest (anterior osteophites). At the core of this method is a new localized shape model that uses concepts of sparsity, dimension reduction, and statistical independence to extract sets of localized modes of deformations specific to each of the vertebrae under investigation. By projection of the shapes onto any specific set of deformation modes (or basis), we obtain low-dimensional features that are most directly related to the pathology of the vertebra of interest. These features are then used as input to a support vector machine classifier to classify the vertebra under investigation as normal or upnormal. Experiments are conducted using contours from digital x-ray images of five vertebrae of lumbar spine. The accuracy of the classification scheme is assessed using the ROC curves. An average specifity of 96.8 % is achieved with a sensitivity of 80 %.

Citation Download Citation

R. Zewail, A. Elsafi, and N. Durdle "Vertebral classification using localized pathology-related shape model", Proc. SPIE 6914, Medical Imaging 2008: Image Processing, 69144P (26 March 2008); https://doi.org/10.1117/12.770314

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Pathology

Spine

X-rays

X-ray imaging

Dimension reduction

Independent component analysis

Feature extraction

Show All Keywords

Keywords/Phrases

Search In:

Publication Years