Automated robust generation of compact 3D statistical shape models

Tomaz Vrtovec; Bostjan Likar; Dejan Tomazevic; Franjo Pernus

doi:10.1117/12.533470

12 May 2004 Automated robust generation of compact 3D statistical shape models

Tomaz Vrtovec, Bostjan Likar, Dejan Tomazevic, Franjo Pernus

Proceedings Volume 5370, Medical Imaging 2004: Image Processing; (2004) https://doi.org/10.1117/12.533470
Event: Medical Imaging 2004, 2004, San Diego, California, United States

Abstract

Ascertaining the detailed shape and spatial arrangement of anatomical structures is important not only within diagnostic settings but also in the areas of planning, simulation, intraoperative navigation, and tracking of pathology. Robust, accurate and efficient automated segmentation of anatomical structures is difficult because of their complexity and inter-patient variability. Furthermore, the position of the patient during image acquisition, the imaging device and protocol, image resolution, and other factors induce additional variations in shape and appearance. Statistical shape models (SSMs) have proven quite successful in capturing structural variability. A possible approach to obtain a 3D SSM is to extract reference voxels by precisely segmenting the structure in one, reference image. The corresponding voxels in other images are determined by registering the reference image to each other image. The SSM obtained in this way describes statistically plausible shape variations over the given population as well as variations due to imperfect registration. In this paper, we present a completely automated method that significantly reduces shape variations induced by imperfect registration, thus allowing a more accurate description of variations. At each iteration, the derived SSM is used for coarse registration, which is further improved by describing finer variations of the structure. The method was tested on 64 lumbar spinal column CT scans, from which 23, 38, 45, 46 and 42 volumes of interest containing vertebra L1, L2, L3, L4 and L5, respectively, were extracted. Separate SSMs were generated for each vertebra. The results show that the method is capable of reducing the variations induced by registration errors.

Citation Download Citation

Tomaz Vrtovec, Bostjan Likar, Dejan Tomazevic, and Franjo Pernus "Automated robust generation of compact 3D statistical shape models", Proc. SPIE 5370, Medical Imaging 2004: Image Processing, (12 May 2004); https://doi.org/10.1117/12.533470

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