A statistical model-based approach for the automatic quantitative analysis of perfusion gated SPECT studies

Sebastian Ordas; Santiago Aguade; Joan Castell; Alejandro F. Frangi

doi:10.1117/12.595897

Translator Disclaimer

14 April 2005 A statistical model-based approach for the automatic quantitative analysis of perfusion gated SPECT studies

Sebastian Ordas, Santiago Aguade, Joan Castell, Alejandro F. Frangi

Author Affiliations +

Proceedings Volume 5746, Medical Imaging 2005: Physiology, Function, and Structure from Medical Images; (2005) https://doi.org/10.1117/12.595897
Event: Medical Imaging, 2005, San Diego, California, United States

Abstract

In this paper we present a statistical model-based approach to three-dimensional (3D) analysis of gated SPECT perfusion studies. By means of a 3D Active Shape Model (3D-ASM) segmentation algorithm, delineations of the endo- and epicardial borders of the left ventricle are obtained, in all temporal phases and image slices of the study. Prior knowledge was captured from a training set of cardiac MRI and SPECT studies, from which geometrical (shape) and grey-level (appearance) statistical models were built. From the fitted shape, a truly 3D representation of the left ventricle, a series of global and regional functional parameters can be assessed. A myocardial center surface representation is built on top of which scalar maps of perfusion, thickness or motion can be depicted. Preliminary results were quite encouraging, suggesting that statistical model-based segmentation may serve as a robust technique for routine use.

Citation Download Citation

Sebastian Ordas, Santiago Aguade, Joan Castell, and Alejandro F. Frangi "A statistical model-based approach for the automatic quantitative analysis of perfusion gated SPECT studies", Proc. SPIE 5746, Medical Imaging 2005: Physiology, Function, and Structure from Medical Images, (14 April 2005); https://doi.org/10.1117/12.595897

ACCESS THE FULL ARTICLE