Detection of LV function abnormality using temporal patterns of normalized wall thickness
© Wael et al; licensee BioMed Central Ltd. 2015
Published: 3 February 2015
Global measures of cardiac function may not reflect subtle wall motion abnormalities. In such cases, assessment of regional cardiac wall motion is required, which is usually evaluated visually with highly subjective results. Automatic analysis techniques can be classified into intensity-based and contour-based methods with the latter category avoiding the image quality-related limitations in the first one. In this study, we present a contour-based technique for detecting wall motion abnormality based on studying the temporal pattern of normalized wall thickness.
The 1-fold leave-one-out method  was applied for training the classifier. Principal-component analysis was applied to find the directions of data variations and for dimensionality reduction. F1-score was used to select the proper number of principal-components. Naïve Bayes' classifier was applied to assign label to each segment (normal or abnormal). Slice abnormality was determined if slice contains two or more abnormal segments.
The proposed method provides automatic assessment of regional myocardial abnormality in a segmental basis for each slice; therefore, it could be a valuable tool for automatic and fast detection of early signs of cardiac dysfunction from conventional untagged cine images.
- Elisseeff A: "Leave-one-out error and stability of learning algorithms with applications". IOS Press. 2002Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.